# n8n Automations

> Behind-the-scenes workflows that connect your apps and AI models so work moves on its own, no manual steps.

Page: https://mburhan.com/tag/n8n

24 case studies in full below.

# Salesforce Marketing Cloud to GoHighLevel Migration & Two-Way CRM Sync

> Automated CRM migration, customer journey rebuilding, and real-time two-way data synchronization.

Section: n8n Automations

Page: https://mburhan.com/case/salesforce-marketing-cloud-to-gohighlevel-migration-two-way

## 1. Overview

My Digital Money runs Cash and IRA investment accounts and grows its list with a free crypto guide. Its customer journeys ran in Salesforce Marketing Cloud, driven by Salesforce leads, accounts and opportunities.

The journeys and contact history were moved into GoHighLevel, and n8n now keeps leads, guide requests and opportunities in step between the two CRMs.

- **Client:** My Digital Money
- **Industry:** Financial Services · Investment Accounts
- **Location:** Encino, California

**Implementation summary:**

- 9 Marketing Cloud journeys documented.
- 18,480 contacts imported in the main migration with 0 errors.
- 3 n8n sync workflows in production.

## 2. The Problem

The objective was to move customer journeys and contact history out of Marketing Cloud and keep Salesforce and GoHighLevel in sync without manual data entry.

The migration involved six main challenges:

1. A hard cut-off: The Marketing Cloud subscription was ending, so journeys, email history and unsubscribes had to be exported before access closed.
2. Journeys wired to Salesforce data: Entry rules and splits read Salesforce fields such as account type (Cash or IRA), funding type and opportunity stage.
3. Two CRMs, no connection: Salesforce stayed the system of record, but GoHighLevel couldn't tell when a lead arrived, asked for the guide or moved through a deal.
4. Silent sync failures: Salesforce outbound messages first hit a test webhook URL, then kept retrying because the reply wasn't in the SOAP format Salesforce expects.
5. Records that had to match: An opportunity could arrive before its contact existed in GoHighLevel, and repeated updates risked creating duplicates.
6. A brand-new sending domain: mail.mydigitalmoney.com had no sending history, so volume had to start at around 1,000 contacts and grow in stages.

## 3. How It Works

The implementation consists of six connected components, covering journey migration, data migration, CRM synchronization and email sending.

Image (https://mburhan.com/uploads/079b39a3e4f932d058085a0b.png)

### Step 1 — Journey Audit and Rebuild

Each Marketing Cloud journey was mapped, including its entry source, criteria, splits, emails and waits. It was then rebuilt as a GoHighLevel workflow with the same branches.

- Splits by account type and funding type carried over as workflow conditions.
- Approval, application, abandonment and first-trade journeys each became their own workflow.
- The lead nurture runs as a multi-day email and SMS sequence after the guide is sent.

**Application submitted, IRA**

Image: Marketing Cloud: New IRA applications split by initial funding type into two application emails. (https://mburhan.com/uploads/925e0edddf0b04c323cdb15f.png)

Image: GoHighLevel: Contact-created trigger with the same funding-type condition, emails and waits. (https://mburhan.com/uploads/660fcd35cee2c40e6f542d3a.png)

**Application submitted — Cash**

Image: Marketing Cloud: One path: application-received email, a one-day wait, then exit. (https://mburhan.com/uploads/1f73341928cf338658572d16.png)

Image: GoHighLevel: Contact-created trigger for unapproved individual (Cash) accounts, then the same email and wait. (https://mburhan.com/uploads/4bb62e90abcdc5bc87a08af9.png)

### Step 2 — Data Migration

Contacts, Marketing Cloud email tracking and unsubscribe data were exported, merged and imported into GoHighLevel.

- Separate imports for unconverted prospects, Cash and IRA account holders, and matched unsubscribes.
- Contact fields created to carry Salesforce data such as account approval, funding type, lead source and guide status.
- The main import processed 18,480 records with no errors.

Image: Bulk import: 18.48K records, 100% success, 0 errors. (https://mburhan.com/uploads/75c26aa9648f7fc7676ed2d0.png)

Image: Contact fields: Salesforce data such as approval, funding type and guide status. (https://mburhan.com/uploads/7562d1012b066ce3d0f7c2bb.png)

### Step 3 — Lead Synchronization

A Salesforce flow sends an outbound message to n8n whenever a lead is saved. n8n reads the full lead and creates or updates the contact in GoHighLevel.

- n8n replies with the SOAP acknowledgement Salesforce requires, so messages aren't retried.
- GoHighLevel custom fields are looked up and mapped before each write.
- A Salesforce "lead created" trigger runs alongside the webhook.

Image: Workflow: webhook or lead trigger → full lead → field mapping → create or update. (https://mburhan.com/uploads/a371036c8fc4cfe007fcbd76.png)

Image: Executions: recent runs succeeded in about a second. (https://mburhan.com/uploads/f05aa0fc4be38666a03a1565.png)

### Step 4 — Guide Requests

When a lead asks for the crypto guide, n8n waits a minute for Salesforce to finish saving, re-reads the lead, sets Send Guide in GoHighLevel and resets it in Salesforce.

- Resetting the flag means a returning lead can request the guide again.
- In GoHighLevel, the Send Guide field starts the guide email and then the nurture sequence.

Image: Marketing Cloud: Splits new leads by campaign before sending the guide emails. (https://mburhan.com/uploads/3048ce17f0bb243e4bb06a4b.png)

Image: GoHighLevel: Send Guide set to true starts the guide email, then the nurture sequence. (https://mburhan.com/uploads/c2f67aa2d06994d813f435f7.png)

Image: GoHighLevel: MDM – Lead Nurture Sequence, shown in three parts (https://mburhan.com/uploads/d1edb27a04eb7053b2a769b6.png)

Image: Workflow: wait → re-read lead → set in GoHighLevel → reset in Salesforce. (https://mburhan.com/uploads/689ad52873bd7f28dce0fd91.png)

Image: Executions: each run includes the one-minute wait (https://mburhan.com/uploads/f2b106b4c567dc57ae33b68e.png)

### Step 5 — Two-Way Opportunity Synchronization

Opportunity changes in Salesforce create or update the matching opportunity in GoHighLevel, and the GoHighLevel ID is written back to Salesforce.

- n8n pulls the opportunity and its account, finds the contact in GoHighLevel and creates it if it's missing.
- Matching uses the Salesforce Opportunity ID stored on the GoHighLevel record, so updates don't create duplicates.
- Salesforce stages map to GoHighLevel pipeline stages, such as Trade – Completed.

Image: Workflow: update or create in GoHighLevel, then write the ID back to Salesforce. (https://mburhan.com/uploads/1e14830b32a8ac41b7cefb52.png)

Image: Executions: successful runs, including batches of updates. (https://mburhan.com/uploads/716b5a20ed235e451b5d6a54.png)

### Step 6 — Sending Domain and Warmup

mail.mydigitalmoney.com was authenticated and set as the sending domain for every email type in GoHighLevel.

- SPF, DKIM, DMARC and MX records verified before sending.
- A contact property controlled who received newsletters during warmup, starting around 1,000 contacts and expanding at each stage.
- Test contacts were removed from production sends so they didn't drag down engagement.

Image: Synced opportunity: stage mapped from Salesforce, with the Salesforce IDs stored (blurred) (https://mburhan.com/uploads/08a942c958ba8a0b1bbc3879.png)

Image: Sending domain: mail.mydigitalmoney.com handles all seven email types. (https://mburhan.com/uploads/42f771534c7087fa4978430a.png)

## 4. The Result

The migration moved the customer journeys and contact history into GoHighLevel while establishing synchronization between Salesforce and GoHighLevel through n8n.

Seven GoHighLevel workflows now cover the guide, nurture, application, approval, abandonment and first-trade journeys. The main contact import processed 18,480 records with 100% success and no errors.

Leads, guide requests and opportunities now sync through n8n. Recent lead sync runs shown in the dashboards completed in about a second, and opportunity records are linked between both systems using their respective IDs.

The new sending domain was authenticated and used for all seven email types, with a staged warmup process.

## 5. Results & Metrics

The following figures are taken from the GoHighLevel and n8n dashboards. They represent the documented implementation results, not independent performance audits.

| Metric | Documented result |
| --- | --- |
| Marketing Cloud journeys documented | 9 |
| GoHighLevel workflows covering customer journeys | 7 |
| Contacts in the main import | 18,480 |
| Main import success rate | 100% |
| Import errors | 0 |
| n8n sync workflows in production | 3 |
| Emails delivered in the last 30 days | 120,362 |
| Email bounce rate | 1.2% |
| Spam complaints | 8 |
| Recent lead sync execution time | About 1 second |

Image (https://mburhan.com/uploads/2bf088211aaa1b1a5527a7dd.png)

## 6. Tech Stack

| Technology | Purpose |
| --- | --- |
| Salesforce Sales Cloud | System of record for leads, accounts and opportunities. |
| Salesforce Marketing Cloud | Original customer journeys and email history. |
| n8n | Self-hosted synchronization between the two CRMs. |
| GoHighLevel | Rebuilt workflows, email, SMS and pipeline management. |
| LeadConnector API | Reads and writes GoHighLevel records. |

## 7. Where This Pattern Applies

- Financial Services: Migrate customer journeys and synchronize investment account information between CRM platforms.
- Insurance Companies: Connect customer records, policy enquiries and application workflows across systems.
- Real Estate Businesses: Synchronize leads, property enquiries and sales opportunities between CRM platforms.
- Marketing Agencies: Migrate email journeys, contact history and nurture sequences while maintaining CRM synchronization.
- Sales Organizations: Keep leads, contacts, opportunities and pipeline stages aligned across multiple CRM systems.

This pattern is useful for organizations moving from one CRM or marketing automation platform to another while maintaining customer history, automated journeys and reliable synchronization between their systems.

---

# One click after the consultation creates a ready-to-work matter

> CASE STUDY · BREWSTER LAW FIRM

Section: n8n Automations

Page: https://mburhan.com/case/one-click-after-the-consultation-creates-a-ready-to-work-mat

## Over View

This project automates Brewster Law Firm’s client intake process using Clio and n8n. When an attorney marks a consultation as **Ready to Engage**, the system automatically creates the matter, assigns the attorney, and sends the client a branded questionnaire with a secure link.

| ~1 min | 3 | 1 click | 10 |
| --- | --- | --- | --- |
| from Ready to Engage to | practice-area groups, | by the attorney; | connected workflows with |
| a created matter | each with its own starting stage | everything after it is automatic | automatic failure alerts |

Once submitted, responses are saved directly in Clio, matter stages are updated, and relevant notifications are sent automatically. The solution reduces manual work, streamlines client onboarding, and keeps matter records organized.

THE PLAN OF ACTION

## We started from a written plan, then built it part by part

Brewster Law Firm's clients arrive through Google and AI search, the website form, the booking page and the phone. Our Plan of Action found that what happened after a lead arrived depended on two people remembering every step. It set out a client journey in five parts, from inquiry to closing, with Clio as the single place where every lead, client and matter lives. Below is each problem the plan found and what we built for it.

| Problem in the plan | What we built | Status |
| --- | --- | --- |
| Form submissions (Part 1) | An instant, branded thank-you with the next step and | Delivered |
| Leads reach Clio Grow, but the lead | booking link, then timed nudges that stop the moment |  |
| gets no thank-you email or next | the lead books. |  |
| step. |  |  |
| Consultation bookings (Part 2) | Confirmation and follow-up emails driven from Clio | Delivered |
| No automatic confirmation email | Grow, alongside Clio's own appointment reminders. |  |
| after a booking. |  |  |
| After the consultation (Parts 2–3) | One click sets the outcome to Ready to Engage. The | Delivered |
| Intake depends on someone | matter is created in the right practice area and first |  |
| remembering each step: convert the | stage, the attorney is assigned and the questionnaire |  |
| lead, create the matter, pick the | goes out. |  |
| stage, send the questionnaire. |  |  |
| Client questionnaires (Part 3) | A branded questionnaire on the firm's own address, with | Delivered |
| Intake forms are collected by hand. | a one-time private link per matter. Answers are saved as a note on the matter. |  |
| Client updates (Part 4) | A plain-language email on every stage change, a "still | Delivered |
| Some clients have to ask for a status | working on it" note when a matter sits too long, and |  |
| update. | reminders when the client owes something. |  |
| Info inbox (Part 1) | Reply tracking pauses automated reminders the moment | Delivered / |
| Replies are manual, usually the next | a client writes back. Reply drafting is the next phase. | next phase |
| morning. |  |  |

| Problem in the plan | What we built | Status |
| --- | --- | --- |
| Ownership and control (Part 0) | Clio stays the single place for leads and matters. | Delivered |
| Tools are only partly set up and | Automations run on a firm-owned, self-hosted n8n |  |
| spread across chat and email. | server, with a daily health report and instant failure alerts. |  |
| Agreements and payments (Part | Kept manual in Clio by the firm's own choice. Billing and | By design |
| 3) | agreements are deliberately not automated. |  |
| Approvals travel by chat; |  |  |
| agreements are hard to track. |  |  |
| Phone, documents, closing (Parts | Voice intake, OneDrive sync and closing and review | Next phase |
| 1, 4, 5) | emails are scoped in the plan for the next phase. |  |
| Smith.ai re-entry, OneDrive folders, |  |  |
| and no closing or review request. |  |  |

## One click by the attorney after the consultation. Everything after it happens on its own.

THE JOURNEY AT A GLANCE

## The plan, drawn out

Image (https://mburhan.com/uploads/b2b6a5624dacd4a23eeeacb7.png)

The five-part client journey. Green is built; dashed grey is planned for the next phase.

Image (https://mburhan.com/uploads/977f32a3fb2516f81c6d6563.png)

From consultation to questionnaire. One click, then the checks, matter, email and stage update happen on their own.

Image (https://mburhan.com/uploads/51e6357bf7337395ece1d612.png)

How the pieces connect. Clio stays the system of record; n8n runs on a firm-owned server.

HOW IT WORKS

## From consultation to questionnaire

1. The outcome is set to Ready to Engage The only manual step. It happens in Clio Grow, where the team already works.

1. The matter is created automatically Formation matters start at "Questionnaire Sent", compliance matters at "Information Gathering", contract matters at "Documents Received". Before it creates anything, the system confirms the client has an email address, matches exactly one contact, and has no open matter. If anything looks unusual, it hands the case to a person instead of guessing.

1. The client receives their questionnaire The email carries a private link tied to that one matter.

1. The client submits it The answers are saved as a note on the matter in Clio, the link closes, and the matter moves to its next stage.

1. Everyone is kept informed The client receives the update for the new stage. The firm receives a short "questionnaire received" notice.

WHAT IT LOOKS LIKE

## Inside Clio and n8n

Image (https://mburhan.com/uploads/78c9aba9c5a20a16b0defe88.png)

The matter-creation workflow. It reads leads from Clio Grow, applies the safety checks, creates the matter, and sends the two emails.

Image (https://mburhan.com/uploads/5599167c7f6de545445814a5.png)

The client questionnaire. Hosted on the firm's own address, in Brewster's branding. Only the description of the matter is required.

Image (https://mburhan.com/uploads/073e9ac9231c6d97ab7b2a96.png)

The matter after the client submitted the questionnaire. It started in the compliance group's first stage and moved to Assessment in Progress on its own. Client details are blurred.

Image (https://mburhan.com/uploads/0ace4ed58b92052fc382b733.png)

The answers land on the matter. The note is saved where the attorney will look for it.

WHAT THE CLIENT RECEIVES

## Emails that read like the firm wrote them

Every email uses the firm's branding and Elissa Brewster Langston's own signature. These are real messages captured from the firm's mailbox during the verification run.

Image (https://mburhan.com/uploads/5c4f45afdcab64dbab17f536.png)

The questionnaire email. Sent the moment the matter is created, with a private link for that one matter.

Image (https://mburhan.com/uploads/d2ee684030e7aca6f848404d.png)

The signature and confidentiality notice match the attorney's own email signature.

Image (https://mburhan.com/uploads/5f292284e5d53781d44a27a5.png)

The stage update. When the questionnaire was submitted and the matter moved, the client received this note on its own. The \[TEST\] tag marks messages sent during verification.

BUILT TO BE TRUSTED

## Guardrails a law firm needs

| The firm's own voice | Every client email is written in the firm's tone and carries the attorney's own |
| --- | --- |
| Per-matter control | signature and confidentiality notice. A single Clio field turns automatic emails on or off for any matter. |
| Private by design | Each questionnaire link works once, for one matter. A reused or invalid link is refused and recorded. |
| Your data stays yours | The workflows run on a dedicated self-hosted server, not a shared automation service. Client records stay in Clio. |

THE RESULT

## The whole journey runs without hand-offs

Run end to end in Brewster's own Clio account, across all three practice-area groups, the system did what the team used to do by hand:

| Matter created | In about a minute, in the correct practice area and first stage, with the attorney |
| --- | --- |
| Questionnaire sent | assigned. A private link arrived in the client's inbox at the same moment, in the firm's branding. |
| Stage moved | Formation, compliance and contract matters each advanced to their next stage when the form came back. |
| Answers on the matter | Saved as a note in Clio, with the team notified, and no one retyping anything. |

RUN YOUR INTAKE THE SAME WAY If your firm uses Clio and still triggers intake by hand, we can map your stages, practice areas and client emails into a workflow like this one. Talk to Growlyze about a free intake review.

---

# Automated Biological Age Engine

> A health and wellbeing company sells a biological ageing assessment. It does not give one “body age”. It shows how each of 16 body systems, from heart and brain to sleep, digestion and skin, seems to be ageing…

Section: n8n Automations, Python Automation

Page: https://mburhan.com/case/automated-biological-age-engine

## Overview

A health and wellbeing company sells a biological ageing assessment. It does not give one “body age”. It shows how each of 16 body systems, from heart and brain to sleep, digestion and skin, seems to be ageing compared with the person’s real age, and what is driving each one.

Each client answers a 150-question form in JotForm and takes a short camera face scan in Upvio. The company had a locked scoring method written up as a specification, but nothing to run it: no way to join each form to its scan, turn the answers into system ages, or produce the personal report it promises.

Now an engine built by Growlyze, with three n8n workflows around it, matches each form to its scan, scores both in code exactly as the specification says, has AI write the report around those numbers, checks every part, and emails the client a private link to their PDF.

## The Problem

Clients pay for a detailed, personal picture of their health. A wrong number or a careless sentence about their body would cost the company their trust.

**Two tools, no shared ID.** The form and the scan arrive separately, at different times and in either order, and the form carries no assessment number. Someone would have to pair every form with its scan by hand.

**The maths has to be exact.** The 150 answers feed 16 body systems, 58 clusters, 28 interaction rules and several indices, with weights that shift when a layer is missing. The same answers must always give the same ages, which an AI model cannot guarantee on its own.

**The scan has to be handled honestly.** A camera reading from one sitting is an estimate. Impossible readings must be left out, not scored, and the report must never name a medical condition.

**The words have to be safe, and every client needs a report.** The text must be in plain British English, must never change or invent a number, and must never make a diagnosis. A weak AI draft cannot be allowed to leave a paying client with nothing.

## How It Works

1. **The form comes in.** An n8n workflow maps the JotForm fields to the engine’s format and sends them on. Each answer must match one of the question’s exact options; an unknown answer stops that assessment and raises an alert instead of guessing. Repeats of the same submission are ignored.
2. **The scan comes in.** A second n8n workflow keeps only completed scans from this assessment’s Upvio link, looks up the client’s email through Upvio’s patient API, and sends the scan to the engine.
3. **Form and scan are matched by email.** Whichever arrives first is stored and waits for the other. When both exist, the engine links them under one assessment ID and queues the report. A scan is never reused for a second report.
4. **Code does the scoring.** The answers and the usable scan readings (blood pressure, resting heart rate, breathing rate and waist-to-height ratio) are scored by the locked method into 16 system ages, burdens and directions. Missing answers are left out, never counted against the client, and readings outside what is physically possible are listed and excluded.
5. **AI writes the words.** The AI gets a pseudonymised summary with no name or contact details and writes the report in parts. It never calculates: the tables and charts are filled by code, and any figure the AI mentions must be one the code already worked out.
6. **Every part is checked.** Checks look for structure, changed or invented numbers, diagnosis claims and hard-to-read sentences. A failing part goes back for repair; minor issues that remain become internal warnings, and any unsafe sentence still there is removed by code, with an alert.
7. **The PDF is built.** The engine adds the client’s first name, builds a branded report of around 45 pages, stores it and creates a private, unguessable link.
8. **The client is emailed.** The engine tells n8n the report is ready, and a third workflow emails the client their link through Gmail, checking first that the event has not already been handled.

All of this runs as its own service, with its own database and four workers, beside the company’s live ADHD assessment without touching it. If the AI service fails, the job is retried and finished parts are kept. The scoring method and the scan mapping are fixed, fingerprinted files, so any change is a new, recorded version rather than a quiet edit.

## Results & Metrics

A form and a scan now find each other by email and become a scored, written, checked report without anyone stepping in. The numbers come only from code, so the same answers always give the same ages, and every accepted assessment ends in a report.

| Metric | Observed Output |
| --- | --- |
| Test reports with the live AI model | 19 |
| Processing time per report (current version) | 3 to 3½ minutes (5 of 5 runs) |
| Report length | 16 body systems, about 45 pages |
| AI drafts checked before publishing | 293 |
| Reports sent with banned or unsafe wording | 0 |
| Deliberate wrong answer and repeated webhook | Stopped with an alert; ignored, no second report |
| Automated tests | 887 passing (100%) |

💡

Metrics come from development and test runs on made-up participants, timed from scored answers to stored PDF. Runs that were restarted after a fix are left out of the timing. Figures from live client traffic may vary.

Image: Workflow: the form intake in n8n. (https://mburhan.com/uploads/35a9fabd2eba1141171348c3.png)

Image: Workflow: the scan intake in n8n, with the email lookup in Upvio. (https://mburhan.com/uploads/113d77a3a273e71e04f70706.png)

Image: Output: the job log of one test assessment with the live AI model, from form to notification in 3½ minutes. IDs shortened. (https://mburhan.com/uploads/115bc8bea36b1529ed37276d.png)

Image: Workflow: the report email in n8n, sent once per client. (https://mburhan.com/uploads/eeb91f90180a1d6cdc9a59dc.png)

Image: End result: two pages of the PDF report for a made-up test participant. Company name blurred. (https://mburhan.com/uploads/1cf6a30ef6b7b99fd7f8dc6f.png)

## Tech Stack

| Technology | Role in the solution |
| --- | --- |
| JotForm | The 150-question assessment form clients fill in |
| Upvio | The camera face scan, and the patient lookup that supplies the email |
| n8n | Three workflows: form intake, scan intake, and the report email |
| Python engine (FastAPI) | Matching, scoring, AI calls, checks, the job queue and the private report links |
| MySQL / MariaDB | Its own database for assessments, scans, scores and the job queue |
| OpenAI | Writes the report text around numbers set by code |
| WeasyPrint | Turns the finished report into a branded PDF |
| Gmail | Sends the client their report link |

The engine runs on the company’s own server, next to its existing ADHD assessment but fully separate from it. Clients use the same kind of form and scan they already know.

## Where This Pattern Applies

This isn’t really about ageing.

**Worked example: a corporate wellness provider.** Employees answer a lifestyle questionnaire and wear a fitness tracker for a week. The two are matched by email, a fixed scoring model turns them into risk areas in code, AI writes each person’s summary around those scores, the text is checked for medical claims, and a private report goes to each employee.

- Clinics: intake form plus device readings
- Insurance: health questionnaire plus checks
- Fitness coaching: check-in plus wearable data
- Property: survey answers plus inspection data
- Education: assessment plus coursework
- Engineering: equipment log plus sensor data

**The shape:** separate inputs are matched to one person, a fixed method scores them in code, AI writes within strict limits, and checks decide what gets sent. What changes from case to case is the questions, the scoring method, the second data source and the list of claims that must never appear.

---

# Automated Health Assessment Engine

> A health and wellbeing company sells an ADHD profile assessment. It is not a diagnosis and not a single “mild to severe” score. It shows how a person’s attention, drive, ideas, order, recovery and body work together,…

Section: n8n Automations, Python Automation

Page: https://mburhan.com/case/automated-health-assessment-engine

## Overview

A health and wellbeing company sells an ADHD profile assessment. It is not a diagnosis and not a single “mild to severe” score. It shows how a person’s attention, drive, ideas, order, recovery and body work together, and it gives strengths and difficulties equal weight.

Each client completes a 72-question form in JotForm and a short camera body scan in Upvio. Before, the answers and the scan sat in those two tools with nothing joining them and nothing turning them into the personal report the company promises.

Now an assessment engine built by Growlyze, with three n8n workflows around it, matches each form to its scan, scores the answers in code, has AI write the report within strict limits, checks every part before publishing, and emails the client a private link to their PDF.

## The Problem

People taking an ADHD assessment are often anxious about what it will say. They have paid for something personal, and a careless report can hurt them.

**Two tools, no shared ID.** The form and the scan arrive separately, at different times and in either order. Upvio’s scan notification does not even include the client’s email, so someone would have to pair every form with its scan by hand.

**The scoring has to be exact.** The 72 answers feed 11 dimensions, scoring bands, nine operating profiles, a seven-day experiment and product-suitability rules. The same answers must always give the same result, which an AI model cannot guarantee on its own.

**The words have to be safe.** The report must use the client’s own words in British English, but never claim a diagnosis or a cause, never mention brain chemicals or medication changes, and never show raw scores or question numbers. A short scan shows how the body was working at that moment; it must not be read as “how much ADHD” someone has.

**Nothing can be lost or mixed up.** Webhooks get sent twice, the AI service can be slow or down, and many clients can finish at once. Every accepted assessment still has to end in one report, sent once, to the right person.

## How It Works

1. **The form comes in.** When a client submits the form, an n8n workflow maps the JotForm fields to the engine’s format, checks it is complete and sends it to the engine. Repeats of the same submission are ignored, and unusable ones (missing answers, no consent) are rejected.
2. **The scan comes in.** A second n8n workflow keeps only completed remote scans, looks up the client’s email through Upvio’s patient API, and sends the scan to the engine.
3. **Form and scan are matched by email.** The engine stores whichever arrives first and waits for the other, with no time limit. When both exist, it links them and queues the report. If someone submits twice, the newest wins, and a scan is never reused for a second report.
4. **Code does the scoring.** The answers are scored against the company’s rules into dimensions, bands, profiles, an experiment and suitability flags. Each scan reading (heart rate variability, recovery, stress, breathing, heart rate, blood pressure and a wellness score) becomes a plain finding and is compared with the questionnaire.
5. **AI writes the words.** The AI gets a pseudonymised summary, with no name or contact details, and is told exactly which findings and which of the client’s own reflections it may use. It writes the questionnaire sections first and the scan section last.
6. **Every part is checked.** The checks look for correct structure, evidence that points to real findings, use of the client’s words, and anything banned: question numbers, raw values, diagnosis or cause claims, medication advice, hype or US spelling. A failing part gets up to two targeted repairs; if it still fails, safe text built from the scores is used, so every client gets a report.
7. **The PDF is built.** The engine adds the client’s name, builds a branded PDF, stores it and creates a private, unguessable link.
8. **The client is emailed.** The engine tells n8n the report is ready, and a third workflow emails the client their link through Gmail, checking first that they have not already been sent it.

All of this runs in a database-backed job queue with several workers. If the AI service fails, the job is retried, and parts that already passed are kept, not written again. The questions, scoring rules and scan ranges live in one versioned configuration file, so a method change, or a whole new assessment, is a configuration edit rather than new code.

## Results & Metrics

The team no longer pairs forms and scans by hand. Each client’s form and scan find each other by email, and a report is scored, written, checked and emailed without anyone stepping in. Nothing reaches a client until it passes the checks, and a temporary AI failure no longer means a missing report.

| Metric | Observed Output |
| --- | --- |
| Test runs with the live AI model | 14 |
| Typical processing time per report | About 1 minute (median) |
| Processing time range | 39 seconds to just over 2 minutes (13 of 14 runs) |
| AI drafts checked before publishing | 66 |
| Reports sent with banned or unsafe wording | 0 |
| Load test | 1,000 submissions, none lost, duplicated or mixed up |
| Automated tests | 110 passing (100%) |

💡

Metrics come from development and test runs, measured from validated input to stored PDF. The one run outside the range waited on automatic retries. Figures from live client traffic may vary.

Image: Workflow: the form intake in n8n. (https://mburhan.com/uploads/15097b0ff45c52c775c254f6.png)

Image: Workflow: the scan intake in n8n, with the email lookup in Upvio. (https://mburhan.com/uploads/0c9a8870e1f6949cf9dd2bab.png)

Image: Output: one engine run for a test participant, from matched form and scan to “ready”. (https://mburhan.com/uploads/51a53421ee744e6d95e3de85.png)

Image: Workflow: the report email in n8n, sent once per client. (https://mburhan.com/uploads/079e08901c2aa7c5e3870712.png)

Image: End result: two pages of the PDF report for a fictional test participant. Logo and company name blurred. (https://mburhan.com/uploads/635714b7d5d5898cde52d2f3.png)

## Tech Stack

| Technology | Role in the solution |
| --- | --- |
| JotForm | The 72-question assessment form clients fill in |
| Upvio | The camera-based body scan, and the patient lookup that supplies the email |
| n8n | Three workflows: form intake, scan intake, and the report email |
| Python engine (FastAPI) | Matching, scoring, AI calls, checks, the job queue and the private report links |
| MySQL / MariaDB | Stores assessments, scans, results and the job queue |
| OpenAI | Writes the report narrative from pre-selected findings |
| WeasyPrint | Turns the finished report into a branded PDF |
| Gmail | Sends the client their report link |

The engine runs on the company’s own server next to its database. Clients keep using the same form and scan, and the team keeps working in the tools it already has.

## Where This Pattern Applies

This isn’t really about ADHD.

**Worked example: a recruitment firm.** A candidate fills in a skills questionnaire and records a task on another platform. The two are matched by email, the questionnaire is scored against the firm’s rubric in code, AI writes a candidate summary from the scores only, the summary is checked for banned claims, and a PDF goes to the hiring manager.

- Clinics: intake form plus test results
- Fitness coaching: check-in plus wearable data
- Education: assessment plus coursework
- Insurance: application plus documents
- HR: engagement survey to team report
- Financial advice: fact-find to review pack

**The shape:** separate inputs are matched to one person, code scores them, AI writes within fixed limits, and checks decide what gets sent. What changes from case to case is the questions, the scoring rules, the second data source and the list of claims that must never appear.

---

# AI-Powered Facebook Content Generation Engine

> An automated content generation system that transforms viral social media posts into original, branded creatives for more than 100 Facebook pages using n8n, Airtable, and AI agents.

Section: n8n Automations

Page: https://mburhan.com/case/ai-powered-facebook-content-generation-engine

### Overview

This solution makes it easier to create and manage content for more than 100 Facebook pages. Airtable is used as the main workspace, while n8n handles the automation. The system collects source content, rewrites captions, recreates images, and adds branding for each page. Virtual assistants can then review the content and schedule it through Post Planner.

### The Problem

Managing content for many Facebook pages takes a lot of time and effort. Virtual assistants have to find trending posts, rewrite captions, recreate images, and add the correct branding for each page. Doing all of this manually makes the process slow and difficult to manage. A single system was needed to bring these tasks together and make content creation easier.

### How It Works

1. **Submit Source Content:** Virtual assistants add a link to a trending post, meme, or news item in Airtable.
2. **Collect Source Content:** n8n retrieves the content and related information from the submitted link.
3. **Rewrite Captions:** AI agents rewrite the original captions to create fresh versions.
4. **Recreate Images:** AI tools process the original images and create new versions.
5. **Add Page Branding:** The system applies the correct templates, logos, and branding for each Facebook page.
6. **Save the Final Content:** The generated captions and images are saved in Airtable for review.
7. **Schedule Posts:** Virtual assistants review the completed content and schedule it through Post Planner.

Image (https://mburhan.com/uploads/0eb0de2085f132c4dc05bed4.png)

### Result

The system brings content collection, AI content creation, and page branding together in one workflow.

- Supports content management for more than 100 Facebook pages.
- Reduces the manual work involved in writing captions and creating images.
- Produces branded content ready for review and scheduling.
- Gives virtual assistants one place to manage their daily content tasks.
- Uses separate n8n workflows to handle different content-generation tasks.

Image (https://mburhan.com/uploads/680619f797cb874f196fe194.png)

Image (https://mburhan.com/uploads/50a564d67f5023966bcf1b96.png)

### Tech Stack

- **n8n:** Automates the workflow and manages content processing.
- **Airtable:** Provides a workspace for submitting source links and reviewing generated content.
- **AI Agents:** Rewrite captions and help create new images.
- **Post Planner:** Used to schedule completed posts.
- **Facebook Pages:** The platforms where the content is published.

### Where This Pattern Applies

This approach can help any organization that manages content for multiple social media accounts.

- **Digital Marketing Agencies:** Create and manage content for different clients.
- **Media Companies:** Turn trending topics into new social media content.
- **Social Media Teams:** Make daily content creation faster and easier.
- **Content Publishers:** Manage content production across multiple pages.
- **Brand Networks:** Keep branding consistent across different social media accounts.

The same approach can be used to collect content, create new versions with AI, add the right branding, and prepare posts for publishing from one central workspace.

---

# AI Voice Agents for Lead Qualification and Customer Support

> An AI voice automation system for Pine Street Code that instantly qualifies new leads, handles customer calls, and connects callers with human representatives when needed.

Section: n8n Automations, AI Agents

Page: https://mburhan.com/case/ai-voice-agents-for-lead-qualification-and-customer-support

### Overview

Pine Street Code wanted to improve lead response times and ensure customer calls were handled even when the team was unavailable. We developed two AI voice agents using ElevenLabs and n8n. Hazel automatically contacts new Meta leads within seconds, asks qualifying questions, and gathers project details. The second agent handles incoming customer calls when the team cannot answer. Both agents integrate with Thrive CRM and Slack for lead tracking and team notifications. A customer-facing application also allows the client to review call recordings and request improvements.

### The Problem

Pine Street Code needed a way to respond to new Meta leads immediately and avoid losing potential customers to competitors. It also needed reliable coverage for incoming calls when staff were unavailable, without changing its existing business phone numbers.

### How It Works

1. **Capture New Leads:** When a prospect submits a Meta lead form, the n8n workflow receives the lead information and initiates the automated calling process.
2. **Immediate AI Follow-Up:** Hazel, the AI lead qualification agent, contacts the prospect within approximately 5–10 seconds of form submission.
3. **Qualify Prospects:** Hazel asks predefined qualifying questions, gathers project requirements, and captures relevant lead details using configured conditions and conversation instructions.
4. **Apply Conversation Guardrails:** Hazel avoids making pricing commitments, disclosing sensitive business information, or making decisions reserved for human representatives.
5. **Notify the Team:** Lead qualification details are sent to Slack so the team can review new opportunities and follow up when human involvement is needed.
6. **Track Leads in the CRM:** The n8n workflows connect with the client's Thrive CRM, maintaining lead information and tracking progress through the pipeline until the opportunity is closed.
7. **Route Existing Business Calls:** The second AI voice agent handles incoming calls through the business's existing phone numbers, allowing Pine Street Code to retain its established numbers.
8. **Activate Customer Support Agent:** When an incoming call rings three times without being answered by the team, the call is routed to the AI customer support agent.
9. **Handle Customer Inquiries:** The support agent responds to customer questions while following strict instructions that prevent unauthorized pricing commitments, disclosure of internal processes, or sharing sensitive business information.
10. **Provide Call Review:** A customer-facing application allows the business to listen to recorded conversations and flag specific calls or interactions for changes and improvements.

Image (https://mburhan.com/uploads/603431b326c2139e55ed68ae.png)

Image (https://mburhan.com/uploads/3fc3d0795b8280cf94b50fd6.png)

### Result

The solution connects lead acquisition, AI-powered qualification, customer support, and CRM tracking into one automated system.

- **Faster lead response:** Hazel initiates calls approximately 5–10 seconds after form submission.
- **Automated qualification:** Prospects are contacted and asked predefined questions to capture project requirements.
- **Human-controlled decisions:** Pricing and other sensitive commitments remain under human control.
- **Improved call coverage:** Incoming calls can be handled by the support agent when the business team does not answer after three rings.
- **Centralized lead tracking:** Lead progress is maintained in Thrive CRM through the final pipeline stage.
- **Client visibility:** The customer-facing application provides access to call recordings and a way to flag interactions for review.

Image (https://mburhan.com/uploads/d92bdba2e07881bca1b0f800.png)

Image (https://mburhan.com/uploads/954f69f4c7d98847fb97196f.png)

### Tech Stack

- **ElevenLabs:** Powers the AI voice agents, conversation handling, and voice interactions.
- **n8n:** Orchestrates lead processing, calling workflows, conditional logic, CRM updates, and notifications.
- **Meta Lead Forms:** Provides incoming prospect information from advertising campaigns.
- **Thrive CRM:** Tracks leads and their progress through the sales pipeline.
- **Slack:** Delivers lead qualification notifications to the business team.
- **Customer-Facing Application:** Allows clients to listen to calls and flag interactions for changes.
- **Telephony Integration:** Routes incoming business calls to the AI customer support agent while retaining existing phone numbers.

### Where This Pattern Applies

The same approach can be adapted for businesses that need faster lead response and reliable phone support.

- **Home Services:** Contact new inquiries immediately and qualify project requirements.
- **Real Estate:** Follow up with property inquiries and identify serious prospects.
- **Marketing Agencies:** Qualify advertising leads and notify sales teams about opportunities.
- **Healthcare:** Handle routine appointment inquiries and route calls when staff are unavailable.
- **Professional Services:** Capture new client requirements and manage incoming service inquiries.
- **Customer Support Teams:** Provide automated call coverage outside busy periods.

The general pattern combines immediate AI-powered lead follow-up, structured qualification, automated customer call handling, and human oversight to help businesses respond faster and manage more conversations efficiently.

---

# From Manual Calendar Checking to Automatic Tutor Matching

> Finding the right tutor for a family used to mean checking the tutor sheet and then each tutor's calendar, one by one. Now a staff member posts the request in Slack, and a ranked list of tutors who are really free arrives in the same thread within minutes.

Section: n8n Automations

Page: https://mburhan.com/case/from-manual-calendar-checking-to-automatic-tutor-matching

## From Manual Calendar Checking to Automatic Tutor Matching

- Browser automation
- n8n automations
- AI agents

**Works with:** n8n, Slack, Skyvern, Google Sheets, TutorBird, Claude

**Before** A tutor sheet and 56 separate calendars to check for every request

**Now** One ranked report in Slack, in about 2 to 4 minutes

## Project overview

A tutoring company places children with tutors. For every family request, one question matters most: *which tutors are free at the times this family needs, and right for this child?*

Answering it means combining two sources: a tutor sheet (who is available now, where they work, which grades, what experience, how many students) and a separate calendar in TutorBird for each of 56 tutors.

Growlyze built a Slack bot that does this check. A staff member posts the request in a fixed numbered format. The bot shortlists the right tutors from the sheet, has a browser automation read their live TutorBird calendars, matches their free time to the family’s windows and session length, and replies in the same thread with a ranked report.

The bot also handles the family’s time zone, applies the company’s rule for how many sessions a week a tutor must cover, and says why a tutor is only a partial match.

## Key results at a glance

| Metric | Before | After |
| --- | --- | --- |
| Time to a match report | 6–8 minutes | 2–4 minutes |
| Calendars read per request | All 56 tutors | Only the shortlisted tutors (24 in the first live test) |
| Tutor information | A sheet plus 56 separate calendars | One ranked report in Slack |
| Name matching | Two name formats: “First Last” in the sheet, “Last, First” in TutorBird | All 34 tutors marked available matched to their TutorBird names |
| Temporary page errors | Re-run by hand | Retried automatically, up to 3 attempts |

Report times come from measured live runs (2:00 and 3:54). The calendar figure comes from the first live test, a remote request with 24 shortlisted tutors. No other numbers are claimed.

## 01. The business challenge

When a family asks for a tutor, speed and accuracy both matter. A family left waiting, or promised a time that can’t be kept, can lose trust quickly. Matching a request to the right tutor involves several connected problems.

### 1.1 Every request has many conditions

A match has to fit the child’s grade, any experience the family needs (such as ADHD or autism), a city or online, the session length (for example 45 or 60 minutes), how many sessions a week, and the family’s free windows in their own time zone. All of these have to be true at once.

### 1.2 The answers live in two places

The tutor sheet says who is available now, where they work, which grades they teach, what experience they have and how many students they already have. Their actual free time sits in a separate TutorBird calendar for each of 56 tutors, which has to be opened through the website.

Image: The TutorBird calendar, where each tutor’s lessons and free time live. Student and tutor names are blurred. (https://mburhan.com/uploads/6b5d88536023df9910eb0385.jpg)

### 1.3 Time zones and names don’t line up

Families write their free times in their own zone (ET, MT, PT and others), while TutorBird shows times in Eastern. The sheet writes names as “First Last”, while TutorBird writes “Last, First”, and some names have a middle name or a two-word surname.

### 1.4 Free time isn’t always usable time

A tutor can be free for 30 minutes inside a family’s window when the session needs 45 or 60. A request for two sessions a week also needs two separate days, not one long gap.

### 1.5 A wrong match is costly

If a tutor looks free and isn’t, the family is promised a time that can’t be kept. A child waits longer for help, and the team has to start the search again.

## 02. The solution: a Slack tutor-matching bot

Growlyze built the bot around the company’s own tutor sheet, TutorBird calendars and Slack channel. It replaces none of them. It connects them, so one Slack message sets the whole check in motion.

### 2.1 A request in Slack

A staff member posts the request as eight numbered lines: student, remote or address, backup option, session length, how often, the family’s free times, grade, and any experience needed. The bot adds a “working on it” reaction and checks every line; if one is missing, it replies in the thread and says which. AI turns the days named in the request, even a range like Monday to Friday, into real dates for the upcoming Sunday-to-Saturday week.

Image: An example request posted in Slack. The sender, student name and address are blurred. (https://mburhan.com/uploads/7f2e31ef27d7678066c1372a.png)

### 2.2 A shortlist from the tutor sheet

The bot reads the tutor sheet and keeps tutors who are marked available now, work in the right city or remote, and teach the child’s grade. AI then checks which of them have the experience the family asked for, matching on meaning (“Autism” and “Autism Spectrum Disorder” count as the same condition) and requiring every condition named. For in-person requests, the bot also works out how far each tutor is from the family.

### 2.3 A live calendar read with Skyvern

The bot sends the dates and the shortlisted tutors, written the way TutorBird writes names, to Skyvern. Skyvern logs into TutorBird, selects those tutors and reads each tutor’s free times for the requested days from that tutor’s own calendar column. The workflow checks every 30 seconds until the read is done, and if TutorBird has a temporary page problem it retries up to two more times.

### 2.4 Matching free time to the family

AI turns the family’s free-text availability into clear day and time windows, including “either day” options. The bot matches sheet names to TutorBird names, converts times from TutorBird’s Eastern time into the family’s time zone, and keeps only the free time that fits a window and is long enough for the session length.

### 2.5 A ranked report in the same thread

The bot then checks how many sessions a week each tutor can cover. Tutors who cover every requested session are full matches; those who cover fewer are partial matches, with the reason and any too-short slots listed; the rest are named as having no availability. The top pick is chosen from the full matches (closest for in-person requests, then fewest students), and remote backups are added when too few in-person tutors fit. The report is posted as a reply to the original message, with a check-mark reaction.

Image: An example report. Student and tutor names, and the logo, are blurred. (https://mburhan.com/uploads/c1644694ab6413577c254253.png)

### 2.6 Clear failure messages

If a step fails, the bot replies in the thread with “Match failed for this request” and the reason, and marks the message with an X. Temporary calendar errors are retried first, so the team only sees an error when every attempt has failed.

## 03. End-to-end workflow

The whole process runs as one n8n workflow, started by the Slack message.

Image: The n8n workflow, from the Slack message to the posted report. (https://mburhan.com/uploads/9ba2e9bdb7552f3e01e98b63.png)

1. Request posted in Slack
2. Dates and fields read from the request
3. Shortlist built from the tutor sheet
4. Calendars read by Skyvern, with automatic retry
5. Free time matched to the family’s windows and session length
6. Weekly frequency checked and tutors sorted
7. Top pick chosen
8. Report posted in the same thread

There is one path from the Slack message to the posted report, and any failure is reported back to the same thread.

## 04. What makes this solution different?

### Built around the company’s own rules

The bot uses the company’s grade bands, experience matching and student counts. When a family accepts “1x or 2x a week”, a full match must cover the higher number, so tutors who can only do one session show as partial.

### Reads only the tutors who could match

The sheet filters run first, and only the shortlisted tutors’ calendars are read. In the first live test that was 24 tutors instead of all 56.

### AI where judgment is needed, fixed rules where precision is needed

AI reads the family’s free-text availability and matches experience and names. Time-zone conversion, overlap, session-length checks and sorting follow fixed rules.

### Partial matches explain themselves

Every partial match says why it falls short (for example, one session of two) and lists any slots that were too short for the session.

### Each tutor is read from their own column

Skyvern takes each tutor’s availability from that tutor’s own calendar column, and the read is set up to check that the number of tutors matches the number of columns.

### Temporary errors fix themselves

If TutorBird has a temporary page problem, the bot retries up to two more times before showing an error.

## 05. Business impact

### Match reports in 2–4 minutes

Measured live runs finished in 2:00 and 3:54. Before tuning, matches had slowed to 6–8 minutes; removing an unused calendar-read step and checking for results every 30 seconds instead of every 2 minutes brought them back down.

### Fewer calendars to read

Only the shortlisted tutors’ calendars are read. In the first live test, that was 24 tutors instead of all 56.

### Less manual re-running

Temporary TutorBird page errors used to mean re-running the request by hand. The bot now retries up to two more times on its own. This has been tested with simulated failures, and the first live retries are still to be observed.

### Clearer shortlists

Each report names the full matches, the partial matches with the reason, and the tutors with no availability, so the team can see why someone was or wasn’t suggested.

### Accuracy checked by hand

In two test requests, all 21 and all 24 tutors in the reports were checked against the raw calendar data and the tutor sheet, and every time range and experience label matched. All 34 tutors marked available in the sheet matched their TutorBird names.

## 06. Technology stack

| Technology | Role in the solution |
| --- | --- |
| n8n | Workflow automation: runs every step, from request to report |
| Slack | Where requests are posted and where the report is replied |
| Skyvern | Browser automation: logs into TutorBird and reads the tutors’ free times |
| TutorBird | The calendar system the team already uses |
| Google Sheets | The tutor roster: availability, location, grades, experience, student count |
| Claude (Anthropic) | AI for dates, free-text availability, experience matching, name matching and ZIP estimates |
| Zippopotam.us | Turns ZIP codes into locations for distance checks |

The technologies support the bot’s main functions: request intake, shortlisting, calendar reading, matching and reporting. The team keeps working in Slack, TutorBird and their own tutor sheet.

## 07. Future enhancements

The following items were identified as future improvements rather than completed capabilities.

### Warn when a tutor can’t be found in TutorBird

Today a tutor the calendar reader can’t find shows up as having no availability. The report should say “not found” instead.

### Refresh and retry inside the calendar reader

Retrying inside the Skyvern run itself would let a temporary page error recover in seconds, instead of starting the whole read again.

### An error message for every failed step

Some steps, such as the AI steps, don’t report a failure to Slack. Adding that would make sure no request is ever left without a reply.

## 08. Where this pattern applies

This isn’t really about tutoring.

**Worked example: a home care agency.** A family asks for a caregiver who speaks Spanish, lives nearby, and is free Tuesday and Thursday afternoons. The sheet shortlists caregivers who are active, certified and in the right area. A browser bot reads only those caregivers’ schedules in the scheduling system, keeps the ones free at the family’s times, and posts a ranked list in Slack.

Where else it fits:

- Salons and spas: stylist to appointment
- Field service: technician to job
- Clinics: provider to patient visit
- Fitness and coaching: trainer to member
- Language services: interpreter to event
- Legal and accounting: advisor to client intake

**The shape:** a request comes in, a sheet shortlists who qualifies, a browser bot reads the live calendars, and a ranked answer lands where the team already works.

**What changes from case to case:** the filters in the sheet, the calendar system being read, and the rules for what counts as a match.

## Conclusion

Matching a family to a tutor depends on a tutor sheet and 56 separate TutorBird calendars, each with its own times, time zones and name formats.

Growlyze connected them with a Slack bot that shortlists tutors from the sheet, reads the shortlisted tutors’ live calendars, converts times to the family’s zone, and replies with a ranked report that explains every partial match.

Measured reports arrive in 2–4 minutes, and temporary calendar errors are retried automatically. The team keeps using Slack, TutorBird and their own tutor sheet.

---

# Salesbeast AI Outbound Sales Machine

> An automated outbound sales system for discovering target companies, enriching decision-maker profiles, evaluating lead quality, and generating personalized email outreach.

Section: n8n Automations

Page: https://mburhan.com/case/salesbeast-ai-outbound-sales-machine

## Overview

Salesbeast is a three-layer AI-powered outbound sales system designed to automate company discovery, decision-maker identification, lead qualification, and personalized email generation.The system collects business information from multiple sources, evaluates potential leads using qualification rules, and uses AI to research decision-makers and prepare tailored email variations.

## The Problem

Finding suitable companies and identifying the right people to contact requires searching multiple platforms, reviewing company websites, checking business information, and researching individual profiles.

After collecting this information, sales teams must manually assess lead quality and write personalized outreach messages. This process can become time-consuming and difficult to manage across large prospect lists.

## How It Works

1. **Collect Company Data:** D7 Networks identifies companies based on industry, location, and company size criteria.
2. **Enrich Business Information:** Google Maps API retrieves ratings, reviews, phone numbers, and business hours. Website scraping collects company information, services, and blog content.
3. **Identify Decision-Makers:** Apollo.io searches for employees based on job titles and retrieves contact details, LinkedIn profiles, and professional background information.
4. **Store Enriched Records:** Company information, decision-maker profiles, and enrichment logs are organized in PostgreSQL.
5. **Qualify Leads:** The qualification layer evaluates company size, Google ratings, review counts, industry, and business activity.
6. **Score and Segment:** Leads receive a score out of 100 and are assigned categories such as Reputation Recovery, Moderate Reputation, High Authority, or industry-specific segments.
7. **Research Decision-Makers:** AI analyzes company information and professional profiles to identify relevant pain points, opportunities, and talking points.
8. **Generate Personalized Emails:** The system creates three email variations for each decision-maker: pain-focused, growth-focused, and authority-focused.
9. **Review Email Quality:** Generated emails are scored, and only emails scoring above 75 are approved. Lower-scoring emails are rejected for regeneration.

Image (https://mburhan.com/uploads/814c3ec401b55c5ce5ac91a3.png)

## Result

The system connects company discovery, contact enrichment, lead qualification, and personalized email generation into one automated process.

It provides:

- Enriched company and decision-maker records.
- Structured lead scoring and segmentation.
- Industry-specific targeting.
- Research-based personalization.
- Three email variations per qualified decision-maker.
- Quality-scored email drafts stored for further use.

The documented design specifies an average email quality score of 88/100. This is a design-document figure, not an independently verified production performance metric.

Image: Figure 1. Company Data Collection & Enrichment (https://mburhan.com/uploads/a2193f4e72128f54291d94ea.png)

Image: Figure 2. Decision-Maker Identification (https://mburhan.com/uploads/7b82c8cdc3072b9a2dc0fe98.png)

Image: Figure 3. AI Lead Qualification & Email Generation (https://mburhan.com/uploads/d5e6ef6b2b97202ac65bbaf2.png)

## Tech Stack

**n8n:** Workflow automation and coordination across the three processing layers.

**D7 Networks:** Company discovery based on search criteria.

**Google Maps API:** Business details, ratings, reviews, and contact information.

**Apollo.io:** Decision-maker identification and professional profile enrichment.

**AI Models:** Lead research, qualification support, and personalized email generation.

**PostgreSQL:** Structured storage for leads, decision-makers, qualification results, and email variants.

## Where This Pattern Applies

- **B2B Sales:** Identify target companies and relevant decision-makers.
- **Marketing Agencies:** Build enriched prospect lists for outreach campaigns.
- **Recruitment:** Research organizations and identify relevant hiring contacts.
- **Business Development:** Discover companies matching specific market criteria.
- **Lead Generation Agencies:** Qualify and segment prospects before outreach.
- **SaaS Companies:** Build targeted outreach campaigns using company and professional profile information.

The general pattern is to combine business data collection, contact enrichment, qualification rules, and AI-assisted personalization into a connected outbound sales workflow.

---

# Case Study Aircall Daily Warm Call

> Automated pipeline that pulls fresh leads from a CRM every morning, sends a personalized heads-up email, and loads a ready-to-dial Aircall campaign before the sales rep even logs in.

Section: n8n Automations

Page: https://mburhan.com/case/case-study-aircall-daily-warm-call

## Overview

This project automates daily warm-lead outreach using n8n, HubSpot, Gmail, and Aircall. Every weekday, it selects up to 30 eligible contacts, sends personalized emails, and adds successfully emailed contacts to a new Aircall campaign. It also updates HubSpot and maintains a backup log to prevent duplicate outreach.

## The Problem

The sales team previously spent time every morning collecting warm leads from HubSpot, sending introductory emails, and manually adding contacts to Aircall campaigns.This repetitive process reduced selling time and created risks such as duplicate outreach, incorrect phone numbers, and calls being made to contacts who had not received the introductory email.

## How It Works

1. **Scheduled Execution:** n8n starts the workflow every weekday at 6:00 AM EST, before the sales team begins its day.
2. **Retrieve HubSpot Contacts:** The workflow collects eligible contacts from a designated HubSpot segment, retrieving up to 500 contact IDs before applying the daily selection limit.
3. **Validate and Select Leads:** It retrieves contact details, normalizes phone numbers into E.164 format, skips contacts already contacted that day, and selects up to 30 contacts.
4. **Prevent Duplicate Outreach:** Selected contacts are marked in HubSpot with the current date. Their details are also saved in a backup data table for auditing.
5. **Send Personalized Emails:** The workflow builds personalized HTML emails using the contact's first name and the sales representative's signature, then sends them through the representative's Gmail account.
6. **Create the Aircall Campaign:** n8n duplicates a master Aircall campaign, adds the current date to its name, and prepares it for the selected contacts.
7. **Synchronize Email and Calls:** Only contacts whose emails were successfully sent are added to the Aircall campaign. Their email activity is also logged on their HubSpot timeline.
8. **Launch the Campaign:** The workflow launches the campaign so the sales representative can open Aircall and start calling without preparing the list manually.

Image (https://mburhan.com/uploads/ca953dceb20268d46c0f95f6.png)

## The Result

The automation connects lead selection, personalized email outreach, and Aircall campaign preparation in one workflow. It removes the need to manually build daily call lists, keeps email and calling activities synchronized, and prevents contacts from being selected repeatedly on the same day.

### Automation Testing Results

| Metric | Result |
| --- | --- |
| Contacts processed per test run | 30 |
| Email delivery success rate | 100% (30/30) |
| Aircall contact loading success | Approximately 93% |
| Duplicate contacts observed | None |
| Daily list and campaign preparation | Automated |

*Note: Results are based on structured testing and dry runs. The remaining Aircall loading failures were linked to malformed phone numbers in legacy CRM records. Call connection rates, email replies, and sales impact have not yet been measured.*

## Tech Stack

- **n8n:** Manages the scheduled workflow, contact selection, email delivery, and campaign creation.
- **HubSpot:** Provides contact records, lead segments, daily outreach flags, and activity logging.
- **Aircall:** Creates and launches daily outbound calling campaigns.
- **Gmail / Google Workspace:** Sends personalized emails from the sales representative's mailbox.
- **Backup Data Table:** Maintains an independent record of selected contacts for auditing and troubleshooting.

## Where This Pattern Applies

- **Sales Teams:** Prepare daily calling lists and coordinate email and phone outreach.
- **B2B Companies:** Follow up with warm leads collected through CRM segments.
- **Lead Generation Agencies:** Automate outreach preparation for multiple sales representatives.
- **Customer Acquisition Teams:** Keep email introductions and follow-up calls organized in one process.

---

# 2–3 weeks 24–48 hours

> CASE STUDY · SUPPLEMENT CONTRACT MANUFACTURING

Section: Custom SaaS Platforms, n8n Automations, AI Agents

Page: https://mburhan.com/case/2-3-weeks-24-48-hours

## Project Overview

Pure NSM manufactures capsule supplements for brands that need their own formulas produced at commercial scale. For every prospective customer, one of the most important questions is straightforward: *How much will this formula cost to manufacture?*

Answering that question, however, involved a complicated process. Quotes were prepared in a large Excel workbook, pricing information was spread across QuickBooks exports and supplier websites, and the manufacturing calculations required detailed knowledge of the business. Only a few experienced team members knew how to complete the workbook correctly.

The result was a quoting process that typically took two to three weeks.

Growlyze replaced this spreadsheet-dependent process with a custom quoting platform that centralizes ingredient and packaging costs, automates manufacturing calculations, supports AI-assisted quote editing, identifies missing supplier prices, and checks formulas for known production risks.

The platform also extends the quoting process to Pure NSM’s website, where prospective customers can receive a ballpark estimate without waiting for a sales call.

Image (https://mburhan.com/uploads/6d9e3e89f2c79f059650057f.png)

Image (https://mburhan.com/uploads/dda4e3d4449b50f7993fc9e7.png)

Image (https://mburhan.com/uploads/a4997711fd0353c56ef08e41.png)

## Key Results at a Glance

| Metric | Before | After |
| --- | --- | --- |
| Quote turnaround | 2–3 weeks | 24–48 hours |
| Raw material records | Distributed across data sources | 7,800+ items in one searchable database |
| Packaging components | Maintained across separate sources | 460+ items centralized |
| Quote preparation | Complex Excel workbook | Guided four-step quote builder |
| Public pricing | Enquiry and waiting process | Three-step ballpark estimator |

The reported improvement reduced quote turnaround from several weeks to one or two days. The centralized database and automated calculations also made pricing information easier to access and the quoting process less dependent on a small number of spreadsheet specialists.

## 01. The Business Challenge

In supplement contract manufacturing, responding quickly with an accurate quote can influence whether a prospective customer chooses a manufacturer. A brand developing a new formula may approach several manufacturers at once. When one manufacturer takes weeks to respond, the opportunity to engage that customer can be lost.

Pure NSM’s previous process created several connected problems.

### 1.1 A Complex Spreadsheet Was the Bottleneck

Every quote depended on a large, multi-tab Excel workbook. Preparing a quote required the user to enter the correct information, follow the existing calculations, and ensure that the different cost components were handled properly.

A mistake in a single cell could affect the final price. Because of this complexity, only a few experienced team members could confidently prepare quotes.

When those team members were busy, other quotes had to wait. The process therefore depended not only on the complexity of the workbook but also on the availability of people who understood how to use it correctly.

### 1.2 Pricing Information Was Scattered Across Multiple Sources

The cost of manufacturing a supplement depends on more than its ingredient quantities. The team also needs current prices for raw materials, packaging, labour, and other relevant production inputs.

Before the new platform, this information was spread across:

- QuickBooks exports.
- Existing ingredient and packaging price lists.
- Historical pricing records.
- External supplier websites.

When an ingredient price was missing, someone had to search supplier websites, compare different package sizes, and manually convert the results into a comparable price per kilogram.

This repetitive research added time to quote preparation and made it harder to maintain a consistent view of costs.

### 1.3 Manufacturing Calculations Required Specialized Knowledge

A supplement quote cannot be calculated through simple multiplication alone. The pricing process must account for how ingredients are used and how manufacturing costs change with formula composition and order size.

The calculations included:

- **Ingredient potency:** Low-potency ingredients can require more raw material to achieve the intended amount in the finished product.
- **Overage adjustments:** Vitamins and minerals may require additional quantities to account for manufacturing requirements.
- **Packaging quantities:** Bottle sizes and order volumes affect the quantities and costs of packaging components.
- **Production labour:** Labour costs vary across 15 different manufacturing steps.
- **Margins and volume tiers:** The final price must reflect the business’s margins and the requested order quantity.

Performing these calculations manually made quote preparation slower and left more room for errors in pricing and margins.

### 1.4 Formula and Production Risks Could Be Identified Too Late

Certain ingredient combinations can create formulation challenges or production difficulties. A formula may also present problems when processed through manufacturing equipment.

When these issues are not identified during quote preparation, the team may discover them only after the quote has already been sent.

Pure NSM therefore needed a way to flag known ingredient and production risks earlier, while the formula was still being reviewed.

### 1.5 Prospective Customers Could Not Get an Immediate Estimate

The existing process required prospective customers to submit an enquiry and wait for the sales team to prepare a quote.

This introduced friction at the beginning of the customer journey. Interested brands could not quickly assess whether a formula was likely to fit their budget, and the sales team spent time responding to enquiries that might not become viable opportunities.

The challenge was to make pricing more accessible without removing the detailed calculations required for manufacturing-specific quotes.

## 02. The Solution: A Custom Manufacturing Quoting Platform

Growlyze built a quoting platform around Pure NSM’s existing pricing logic, data, and operational requirements.

Rather than simply transferring the spreadsheet into another interface, the platform separates the process into connected components: centralized cost management, guided formula entry, automated calculations, AI-assisted changes, supplier price discovery, formula checks, quote generation, and public price estimation.

Each component addresses a specific weakness in the original process.

### 2.1 Centralized Database for Costs and Materials

The first step was to bring the information required for quoting into one searchable database.

The platform consolidates:

- More than 7,800 raw material records.
- More than 460 packaging components.
- Labour rates.
- Supplier contacts.
- QuickBooks pricing information.

The team can search for items, edit records, and perform bulk price updates rather than repeatedly consulting different files and websites.

Each price also records its source, such as the internal database, QuickBooks, a supplier website, or a manual entry. This helps the team understand where a cost originated when reviewing a quote.

The centralized database provides the foundation for the platform’s other features: calculations can use the available cost data, while missing prices can be identified for further research instead of being treated as known values.

### 2.2 Guided Quote Builder With Automated Calculations

The quote builder organizes the preparation process into four steps: Formula, Packaging, Labour, and Summary.

This structure makes the workflow easier to follow and separates the main areas that contribute to the final price.

**Step 1 — Formula**

The team enters the ingredients and quantities required for the proposed supplement formula. The platform applies the relevant potency and overage adjustments as part of its calculation process.

**Step 2 — Packaging**

The packaging step accounts for the selected bottle sizes and the required packaging quantities, allowing packaging costs to be incorporated into the quote.

**Step 3 — Labour**

The platform estimates labour hours and costs across 15 production steps. This allows the labour component to be included in the overall calculation rather than handled separately through manual spreadsheet work.

**Step 4 — Summary**

The system combines the calculated costs, applies Pure NSM’s margins, and produces six price tiers. These cover two bottle sizes at three different order volumes.

The calculation logic was checked against a real historical quote, and the resulting price matched to within ten cents. This provided a practical validation point against the company’s existing quoting process.

### 2.3 AI Assistant for Natural-Language Quote Editing

The platform includes an AI assistant that allows team members to request quote changes in plain English.

Instead of manually locating every relevant field, a user can enter an instruction such as:

- “Add 200mg magnesium glycinate.”
- “Set bottle cost to 75 cents.”

The assistant interprets the instruction and proposes the requested changes. The changes are not applied automatically: a team member must approve them first.

This approval step keeps the user involved in pricing decisions and reduces the risk of unintended changes to a quote.

The assistant also follows an important pricing safeguard: **it is not allowed to guess missing prices.**

If a required cost is unavailable, the assistant identifies the missing information and directs the user to the Price Finder. This ensures that an unknown price is treated as a gap to resolve rather than an opportunity for the AI to invent a value.

### 2.4 Price Finder for Supplier Research

Ingredient prices may be missing from the internal database, particularly when the team needs to quote an ingredient that has not previously been used or priced.

The Price Finder streamlines this process.

The team can paste in up to 50 ingredients at once. The system first checks Pure NSM’s own database and then searches supplier websites when additional pricing information is required.

The results are processed so that different package sizes can be compared on a price-per-kilogram basis. The system also assesses how closely each result matches the requested ingredient and presents the available options side by side.

After reviewing the options, the team can select a price and add it to the quote with one click.

This approach reduces repetitive supplier research and brings price discovery into the quoting workflow itself.

### 2.5 Formula Safety and Production-Risk Checks

The platform checks ingredient changes against known problematic combinations and production risks.

These checks are integrated into the quote-building process so that potential issues can be highlighted while the team is still preparing the quote.

This gives the team an opportunity to review known concerns before the quote is sent, rather than relying entirely on someone noticing the issue during a later review.

The purpose is to surface potential risks earlier; the checks do not replace the need for appropriate formulation and manufacturing review.

### 2.6 One-Click Quote Document Generation

Once a quote is ready, the team can click Send to generate a formatted quote document.

The document includes:

- All six pricing tiers.
- The associated cost breakdown.
- The 50% deposit amount.

This reduces the manual work required to assemble the final quotation after the calculations have been completed.

Instead of preparing the calculations in one place and formatting the customer-facing document separately, the platform connects quote preparation with document generation.

### 2.7 Website Estimator for Prospective Customers

The same pricing engine also powers a public-facing estimator on Pure NSM’s website.

Prospective customers enter their formula and quantity through a three-step process. They then receive a ballpark price per bottle on screen and by email, without needing an account or a sales call.

The estimator also sends the customer’s details and formula information to the sales team.

This creates a direct connection between the pricing platform and lead generation. Prospective customers can explore indicative pricing independently, while the sales team receives information about the formula they are interested in.

The website estimator provides a ballpark price; it should not be confused with a fully reviewed manufacturing quotation.

## 03. End-to-End Workflow

The platform connects its internal pricing process and public-facing estimator through the same underlying pricing logic.

Image (https://mburhan.com/uploads/8a0d4fc58f41b6a6bd1809b0.png)

**Public-facing path:**

Image (https://mburhan.com/uploads/ddfdca6bdfb580ab0ac80981.png)

These paths address two related goals: helping the internal team prepare quotes faster and giving prospective customers a quicker way to understand indicative pricing.

## 04. What Makes This Solution Different?

### Built Around Pure NSM’s Manufacturing Logic

The platform was designed for the pricing requirements of supplement manufacturing, including potency, overage, packaging, production labour, and order-volume tiers.

It uses Pure NSM’s own pricing logic rather than relying on a generic quoting process.

### AI With Guardrails

The AI assistant proposes changes for human approval and does not invent missing prices. This keeps pricing decisions under team control while still providing the convenience of natural-language instructions.

### Supplier Research Integrated With Quoting

The Price Finder connects missing ingredient costs with supplier research and price normalization. The team can compare potential matches and bring a selected price directly into the quote.

### Cost Sources Remain Visible

Prices retain source information, helping the team review where costs came from instead of treating every figure as an unexplained spreadsheet value.

### Internal Quoting and Public Estimation Share One Pricing Engine

The website estimator extends the same pricing system to prospective customers. This connects internal quoting operations with a self-service route for generating enquiries.

## 05. Business Impact

### Quote Turnaround Reduced to 24–48 Hours

The reported turnaround improved from two to three weeks to 24–48 hours. Pure NSM can now respond to prospective customers much faster, which is particularly important when brands are comparing multiple contract manufacturers.

### Less Dependence on Spreadsheet Specialists

The guided workflow and automated calculations make quote preparation less dependent on a few experienced spreadsheet users. The process is organized into defined stages that other team members can follow.

### More Consistent and Traceable Pricing

The platform brings cost records into one searchable location, records price sources, and applies the manufacturing calculations consistently. Its calculation logic was also checked against a historical quote, with a difference of no more than ten cents.

### Earlier Identification of Formula Risks

Known ingredient and production concerns can be surfaced during quote preparation. This gives the team more opportunity to review potential issues before sending a quotation.

### Faster Access to Supplier Pricing

The Price Finder reduces the need to research and normalize supplier prices manually for every missing ingredient. It supports batches of up to 50 ingredients and presents potential matches for review.

### Improved Customer Access to Indicative Pricing

The website estimator lets prospective customers obtain a ballpark estimate without first arranging a sales call. The sales team receives their details and the formula they are interested in, creating a more direct path from pricing exploration to follow-up.

## 06. Technology Stack

| Technology | Role in the Solution |
| --- | --- |
| React | Frontend application interface |
| Node.js | Application backend |
| OpenAI | AI-assisted quote editing |
| Firecrawl | Supplier website research |
| n8n | Workflow automation |
| Google Sheets | Supporting data workflows |
| QuickBooks data | Existing pricing information used by the platform |

The technologies support the platform’s main functions: cost-data management, quote calculations, AI-assisted changes, supplier price discovery, and connected workflows.

## 07. Future Enhancements

The following items were identified as future improvements rather than completed capabilities.

**Quote History**

Save quotes and their history directly in the platform so previous pricing decisions can be reviewed more easily.

**CRM Integration**

Feed winning quotes back into the CRM to track win rates and margins by customer. This would help connect quoting activity with subsequent sales outcomes.

**Additional Formula and Packaging Guidance**

Extend the safety-check process with capsule-size and bottle-fit recommendations to provide additional guidance during quote preparation.

## Conclusion

Pure NSM’s original quoting process relied on a complex spreadsheet, dispersed pricing information, and specialist knowledge. Preparing a quote could take two to three weeks, while prospective customers had to wait for the sales team to provide an estimate.

Growlyze replaced that process with a custom quoting platform that centralizes more than 7,800 raw materials and 460+ packaging components, automates manufacturing calculations, supports AI-assisted editing with human approval, helps locate supplier prices, flags known formula risks, and generates formatted quote documents.

The same pricing engine also powers a website estimator, allowing prospective customers to obtain a ballpark price without a sales call.

The reported result is a reduction in quote turnaround from **two to three weeks to 24–48 hours**, supported by a more structured and traceable pricing proces

---

# PeakNIL Athlete Content Automation

> Two n8n workflows that find athlete news, draft the post, and publish it to Instagram once a person approves.

Section: n8n Automations

Page: https://mburhan.com/case/peaknil-athlete-content-automation

## Project Overview

PeakNIL connects college athletes with brands for Name, Image, and Likeness (NIL) opportunities. To attract brand partnerships and maintain a consistent social media presence, the team needs a steady stream of posts highlighting athlete achievements across 28 sports.

This project uses two connected n8n workflows to automate the process. The first monitors athlete news, evaluates stories, drafts captions, and sends content for review. The second detects the manager’s email response, generates a branded graphic, and publishes the post to Instagram after approval.

Human approval remains part of the publishing process, helping ensure that the final content aligns with PeakNIL’s brand and content standards.

**Technology stack:** n8n, Claude, Google Sheets, Gmail, Placid, Instagram.

## The Business Problem: Scaling Athlete Coverage

PeakNIL manages hundreds of athletes across 28 sports. Finding relevant stories and turning them into social media content requires ongoing research, writing, image selection, graphic design, and publishing.

The main challenges include:

- **Time-consuming research:** Staff must regularly check news about a large athlete roster.
- **Manual content production:** Captions, headlines, photos, and branded graphics require several separate steps.
- **Inconsistent coverage:** Manual processes make it difficult to maintain a regular posting schedule across all athletes.
- **Incomplete athlete records:** Missing information can make it harder to identify and track athletes.
- **Brand safety:** Posts must focus on positive, relevant achievements and avoid injuries, losses, and controversies.
- **Publishing control:** Content must be reviewed and approved before it appears on Instagram.

The objective was to reduce repetitive work while improving consistency and retaining human control over publication.

## The Solution

Two connected automations handle the work, and a person approves the result.

### Workflow 1: Athlete Monitor (finds and drafts)

- Runs every weekday and picks athletes from the roster sheet. It skips anyone with missing data or scanned in the last 3 days.
- Searches for recent news (last 3 days) from trusted sports sources.
- Claude decides whether the story is worth posting, assigns a content category, and writes the caption and a short headline.
- Finds a clean photo of the athlete and emails the draft to the content manager for review.

Image: Workflow 1, Athlete Monitor: a schedule trigger reads the roster from Google Sheets, searches for news, has Claude draft the post, and emails it to the content manager through Gmail. (https://mburhan.com/uploads/264b42eea4bba61338942df8.png)

### Workflow 2: Reply Detection (approves and publishes)

- The manager replies to the email with an image number or “approve”.
- The workflow reads the reply, builds a branded graphic with the chosen photo and headline, and sends it back for a final check.
- Once approved, it publishes the post to Instagram and updates the tracking sheet.

Image: Workflow 2, Reply Detection: a Gmail trigger reads the manager’s reply, then one branch builds the graphic with Placid and another publishes to Instagram after approval (https://mburhan.com/uploads/efb8246af99dc1a47b5cb476.png)

## How the System Works

### Step 1: Read the Athlete Roster

The Athlete Monitor workflow retrieves athlete records from Google Sheets and checks whether each record contains the information needed for monitoring.

Athletes with incomplete data or who were scanned within the previous three days are skipped.

### Step 2: Discover Recent News

For eligible athletes, the workflow searches trusted sports sources for relevant news published within the last three days.

This creates a pool of recent stories that can be evaluated for potential social media content.

### Step 3: Evaluate and Draft Content

Claude evaluates each story for suitability, assigns a content category, and generates a short headline and caption.

The content must follow PeakNIL’s positive, brand-safe approach, avoiding injuries, losses, and controversies.

### Step 4: Send the Draft for Review

The workflow finds a suitable athlete photo and sends the proposed content to the content manager by email.

The manager can review the story and select an image or provide an approval response.

### Step 5: Generate the Branded Graphic

The Reply Detection workflow reads the manager’s response and uses Placid to create a branded graphic from the selected photo and headline.

The graphic is sent back for a final check before publishing proceeds.

### Step 6: Publish and Update Tracking

After approval, the workflow publishes the post to Instagram and updates the athlete content tracking sheet in Google Sheets.

This closes the loop between discovery, content creation, review, and publication.

## 05. Workflow Architecture

### Workflow 1: Athlete Monitor

**Weekday schedule → Google Sheets athlete roster → Eligibility checks → Recent news search → Claude story evaluation and caption writing → Athlete photo selection → Gmail review email**

### Workflow 2: Reply Detection

Image (https://mburhan.com/uploads/863794ec55159c75644431ea.png)

The first workflow prepares content for review. The second uses the manager’s response to continue the process and publish only after approval.

## Content Rules and Human Approval

The automation is designed to maintain a positive and brand-safe social media presence.

### Story Selection

Claude evaluates whether a news story is worth posting and assigns it a content category before generating the draft.

### Brand Safety

The content rules exclude stories focused on injuries, losses, or controversies. This helps keep the proposed content aligned with the project's positive athlete-celebration approach.

### Human Review

The content manager receives the draft by email and provides a response. The workflow then generates the branded graphic and sends it for a final check.

**Publishing remains approval-gated:** the workflow is designed to publish to Instagram only after the required human approval.

## 07. Integrations and Their Roles

| Integration | Responsibility |
| --- | --- |
| n8n | Orchestrates both workflows and connects the services. |
| Google Sheets | Stores the athlete roster and tracks content activity. |
| Claude | Evaluates news stories and drafts headlines and captions. |
| Gmail | Delivers draft emails and detects manager responses. |
| Placid | Generates branded social media graphics. |
| Instagram | Receives approved social media posts. |

Each service has a defined role, allowing the workflow to move content between research, writing, review, design, and publishing.

## 08. Results and Business Value

The project is designed to reduce the manual effort required to create athlete-focused social media content while improving consistency across the roster.

### Delivered Capabilities

- Automated weekday monitoring of athlete records.
- Recent-news discovery using a three-day lookback.
- AI-assisted story evaluation, categorization, and caption writing.
- Photo selection and email-based content review.
- Automated branded graphic generation.
- Reply-driven approval processing.
- Instagram publishing after approval.
- Google Sheets tracking updates.
- Checks for incomplete athlete records and recently scanned athletes.

### Expected Business Value

- Less manual time spent researching athletes and drafting posts.
- More consistent coverage across sports and athletes.
- Faster preparation of social media content.
- A repeatable review and publishing process.
- Better visibility into athlete records that need attention.

The project description does not provide measured time savings, publishing volume, engagement improvements, or brand-partnership conversion figures. These should be added after reviewing actual workflow and Instagram performance data.

## Key Takeaways

- **Separate discovery from publishing:** The two-workflow design keeps research and drafting distinct from approval and publication.
- **Use AI to reduce repetitive work:** Claude handles story evaluation, categorization, headlines, and captions.
- **Keep people in control:** Human review remains part of the publishing process.
- **Standardize visual content:** Placid generates branded graphics from the selected photo and headline.
- **Maintain centralized tracking:** Google Sheets supports athlete records and content tracking.
- **Improve coverage consistency:** Automated monitoring provides a repeatable process for discovering stories across a large roster.

## Future Development

- Measure time spent per post before and after automation.
- Track the number of stories discovered, drafts reviewed, and posts published.
- Measure engagement by sport, athlete, and content category.
- Add monitoring for failed news searches, missing photos, and incomplete athlete records.
- Evaluate the quality of AI story selection and caption drafts through reviewer feedback.
- Add reporting on publishing frequency and approval turnaround time.

## Final Outcome

PeakNIL Athlete Content Automation connects athlete news discovery, AI-assisted drafting, email-based review, branded graphic creation, and Instagram publishing in two n8n workflows.

The system is designed to help PeakNIL maintain a consistent flow of athlete-focused content while reducing repetitive production work. By keeping human approval in the publishing process, it balances automation with editorial oversight and brand safety.

---

# AI Post Rewrites and CTA Rotation for a Content Team

> Two n8n workflows that rewrite a social post when a reviewer ticks Redo, and add the best call-to-action to each thread from a rotating library.

Section: n8n Automations

Page: https://mburhan.com/case/n8n-social-post-redo-and-cta-rotation

## Project Overview

This project uses two n8n workflows to improve a content team's social media review and publishing preparation process. Drafts for X, LinkedIn, and Threads are stored in Airtable, where reviewers can request revisions and manage content before approval.

The first workflow, **Redo**, rewrites posts when a reviewer provides feedback and checks the Redo box. Claude applies the requested changes while preserving the existing content that does not need revision.

The second workflow, **CTA Auto-Insert**, selects suitable calls to action from an Airtable library, arranges them within a social thread, and tracks usage to help balance CTA rotation.

Both workflows use a shared Claude sub-workflow and retrieve their writing rules, playbook, and CTA configuration from Airtable. This allows the team to adjust content standards and CTA selection without editing the main n8n workflows.

**Technology stack:** n8n, Airtable, Claude, X, LinkedIn, Threads.

## 01. The Business Problem: Revisions and CTAs Require Consistent Handling

Social media drafts often need several rounds of review before they are ready. Reviewers may request specific changes, but applying feedback across multiple platform formats can be time-consuming and may unintentionally alter content that was already correct.

Calls to action create another challenge. A team may maintain several CTAs for different topics, campaigns, and audiences. Selecting them manually can lead to repetitive links, inconsistent placement, and excessive use of the same CTA.

The main challenges were:

- **Manual revision work:** Reviewer feedback must be applied consistently across multiple social platforms.
- **Content consistency:** Existing facts, approved wording, and account-specific voice rules must be preserved.
- **Platform-specific formatting:** X, LinkedIn, and Threads require different content lengths and formatting.
- **CTA repetition:** Frequently used CTAs can become overused if rotation is not managed.
- **Changing content rules:** Writing standards and CTA preferences need to remain editable without modifying workflow logic.
- **Duplicate processing:** A card should not be rewritten or assigned a CTA repeatedly after its task has been completed.

The objective was to automate these repetitive tasks while retaining reviewer control and clear records of each content change.

## 02. The Solution: Two Connected Content Workflows

### Workflow 1: Redo — AI-Powered Post Rewriting

The Redo workflow checks Airtable for cards where the reviewer has selected the Redo checkbox and provided source text.

A dispatcher can start the workflow, while an hourly scheduled run catches requests that may have been missed.

The workflow retrieves the current Post Rules and Craft Playbook from Airtable and sends Claude the original source, previous draft, and reviewer feedback. Claude produces revised versions for X, LinkedIn, and Threads, following the required platform formats and account voice.

A cleanup step removes unwanted hedging language, applies image-credit rules, and prepares the output for review. n8n then writes the revised posts back to Airtable, clears the Redo checkbox, and changes the card status to `Pending Review`.

The original feedback remains on the card for reference.

### Workflow 2: CTA Auto-Insert — Automated CTA Selection and Rotation

The CTA workflow searches Airtable every two minutes for draft cards that do not yet have a linked CTA.

It retrieves active CTA records, filters out those that have reached their usage caps, and ranks the remaining options using weight, conversions, and previous usage.

Claude receives the post, its source material, and the ranked CTA list. It selects a primary CTA and, when appropriate, a different secondary CTA based on topic relevance and each CTA's usage notes.

The workflow builds the thread with the primary CTA as the first reply and the secondary CTA at the end. It then saves the thread, links the primary CTA record to the card, and increments the usage counters.

**Operational status:** The Redo workflow is live. CTA Auto-Insert is switched off in the current version because running every two minutes would use approximately 21,600 executions per month.

## 03. Project Snapshot

| Item | Detail |
| --- | --- |
| Platform | n8n |
| Content database | Airtable |
| AI writing and selection | Claude |
| Supported platforms | X, LinkedIn, Threads |
| Main workflows | Redo and CTA Auto-Insert |
| Shared component | Claude writing sub-workflow |
| Writing configuration | Post Rules and Craft Playbook in Airtable |
| CTA configuration | Airtable CTA library |
| Redo trigger | Dispatcher plus hourly scheduled check |
| CTA trigger | Every two minutes in the configured workflow |
| Redo status | Live |
| CTA Auto-Insert status | Switched off in the current version |
| Duplicate prevention | Redo checkbox reset; CTA record linked to card |

## 04. How Workflow 1 Works: Redo

### Step 1: Identify Revision Requests

The workflow searches Airtable for cards with Redo enabled and source text available.

Requests can be started through a dispatcher, while the hourly run catches cards that were not processed by the initial trigger.

### Step 2: Load Current Writing Rules

The workflow retrieves the Post Rules and Craft Playbook from a settings record in Airtable.

These settings define the account's writing style and content requirements, allowing the team to update its standards centrally.

### Step 3: Send Context and Feedback to Claude

Claude receives three inputs:

- The original source material.
- The previous draft.
- The reviewer's feedback.

The model is instructed to apply the requested changes while retaining content that the reviewer did not ask to change.

It returns three platform-specific drafts:

| Platform | Required format |
| --- | --- |
| X | Revised X post |
| LinkedIn | 600–1,100 characters |
| Threads | Approximately 420 characters, with a maximum of 500, ending on a complete sentence |

### Step 4: Clean and Validate the Drafts

A cleanup step removes hedging lines such as “the source does not establish…” and applies the required image-credit formatting.

The “shared by a publisher” credit line is retained only for posts containing an image and is moved to the final line.

The prompt also enforces factual accuracy, quotation requirements, and account-specific voice rules.

### Step 5: Update Airtable for Review

n8n writes all three revised drafts back to the Airtable card, unticks Redo, and sets the status to `Pending Review`.

Clearing the checkbox prevents the same revision request from being processed repeatedly, while retaining the feedback provides an audit trail for the reviewer.

## 05. How Workflow 2 Works: CTA Auto-Insert

### Step 1: Find Drafts Without a CTA

The workflow checks Airtable every two minutes for draft cards that do not have a CTA linked yet.

Once a CTA is linked, the card no longer matches the search criteria, preventing it from being processed again.

### Step 2: Retrieve the CTA Library

The workflow reads active CTA records from Airtable. Each record can include:

- CTA text and link.
- Category.
- Weight.
- Times used.
- Conversions.
- Weekly usage cap.
- A “use when” note describing the appropriate context.

This makes CTA selection configurable without changing the n8n workflow.

### Step 3: Filter and Rank CTA Options

A code step removes CTAs that have reached their usage caps and scores the remaining candidates.

The ranking formula uses the following rules:

- Each CTA's weight counts double.
- Each conversion adds half a point.
- Each previous use subtracts one point.

The remaining candidates are sorted from highest to lowest score.

This ranking helps balance relevance, performance, and usage frequency. Claude still considers the topic and each CTA's usage notes before making its final selection.

### Step 4: Ask Claude to Select the CTAs

Claude receives the post, its source material, and the ranked CTA options.

It selects a primary CTA that fits the topic and a different secondary CTA when appropriate. When two candidates are equally suitable, the higher-ranked option is preferred.

The generated reply follows the account's voice, and CTA links must remain exactly as stored in Airtable.

### Step 5: Assemble and Save the Thread

The workflow builds the thread in this order:

1. Lead post.
2. Primary CTA as the first reply.
3. Remaining thread content.
4. Secondary CTA at the end, when selected.

Each part is numbered sequentially, such as `1/`, `2/`, and so on.

n8n saves the completed thread to Airtable and links the primary CTA record to the card. It then increments the Times Used counter for the primary CTA and, when a secondary CTA is included, the secondary CTA as well.

## 06. Workflow Architecture

### Redo workflow

Image (https://mburhan.com/uploads/2b20d2ee118a67e5f7bdb10a.png)

### CTA Auto-Insert workflow

Image (https://mburhan.com/uploads/a69d46969575084a02393736.png)

Both workflows share the same Airtable content base and Claude writing sub-workflow. This reduces duplicated writing logic and helps maintain consistent content-generation behavior.

### n8n Work Flow

Image: Together, the workflows demonstrate how n8n, Airtable, and Claude can automate repetitive content operations while retaining configurable rules, traceable reviewer feedback, and human approval. (https://mburhan.com/uploads/17853f24b5fb8b0133866532.png)

## 07. AI Writing Rules and Content Safeguards

The Redo workflow is designed to apply feedback without introducing unsupported information or changing unrelated content.

### Factual Accuracy

- Use facts contained in the source material or previous draft.
- Never invent numbers, dates, figures, or view counts.
- Return the complete revised post rather than a list of suggested changes.

### Quotation Handling

If the reviewer requests a quote, the publisher's own words must remain verbatim and enclosed in quotation marks. Any interpretation or analysis must be written separately after the quotation.

### Voice and Formatting

The workflow follows the account's voice rules, including a list of banned words and AI-sounding phrases.

It also applies platform-specific length requirements and cleans up unwanted hedging language before writing the drafts back to Airtable.

### CTA Controls

CTA Auto-Insert filters out options that have reached their usage caps, uses a ranking system to guide selection, and preserves stored CTA links exactly.

The primary and secondary CTAs are intended to be different, helping reduce repetitive calls to action within the same thread.

These controls guide the output but do not replace a human review of factual accuracy, tone, and suitability.

## 08. Airtable as the Content Control Center

Airtable stores the drafts, review feedback, writing settings, and CTA library used by both workflows.

| Component | Purpose |
| --- | --- |
| Content cards | Store source material and platform-specific drafts. |
| Redo checkbox | Signals that a reviewer wants the drafts rewritten. |
| Feedback field | Captures the changes requested by the reviewer. |
| Post Rules | Defines the account's content requirements. |
| Craft Playbook | Provides the broader writing guidance. |
| CTA library | Stores CTA text, links, categories, scores, usage caps, conversions, and context notes. |
| Linked CTA field | Identifies the CTA assigned to a card and prevents repeated processing. |
| Times Used counter | Tracks CTA usage for rotation. |

Centralizing these settings allows the team to modify its writing guidance and CTA options without editing the underlying workflow.

## 09. Results and Operational Status

The project description confirms that the Redo workflow is live and the CTA workflow is configured but switched off in the current version.

No measured figures were provided for hours saved, revision turnaround, approval rates, engagement improvements, or CTA conversion uplift.

### Implemented Capabilities

- Reviewer-triggered AI rewriting.
- Hourly recovery checks for missed Redo requests.
- Platform-specific drafts for X, LinkedIn, and Threads.
- Centralized writing rules and playbook settings.
- Preservation of source facts and reviewer feedback.
- Cleanup of unwanted hedging language.
- Image-credit formatting rules.
- CTA filtering and weighted ranking.
- Primary and secondary CTA placement.
- Airtable-based usage tracking and duplicate prevention.

### Important Operational Consideration

The CTA workflow is switched off because its two-minute schedule would generate approximately 21,600 executions per month. Before enabling it, the team should evaluate a less frequent schedule or event-based triggering to reduce execution consumption.

### Metrics to Measure Next

- Average time from reviewer feedback to revised drafts.
- Percentage of rewritten drafts approved without another revision.
- Frequency of unsupported facts or formatting corrections.
- CTA processing volume and execution consumption.
- CTA usage distribution and weekly-cap enforcement.
- Click-through and conversion performance of approved CTAs.

## 10. Key Takeaways

- **Automate targeted revisions:** Reviewers can request changes through Airtable without manually rewriting every platform version.
- **Preserve existing content:** Claude receives the source, previous draft, and specific feedback so unrelated content can remain unchanged.
- **Centralize writing standards:** Post Rules and the Craft Playbook are editable in Airtable.
- **Make CTA rotation data-driven:** Ranking uses weight, conversions, and usage history, while Claude considers contextual relevance.
- **Prevent duplicate processing:** Clearing Redo and linking a CTA give each workflow a clear completion condition.
- **Consider execution costs:** Frequent scheduled checks should be balanced against operational needs and n8n execution limits.
- **Keep human review in the loop:** Rewritten drafts return to `Pending Review` rather than being published automatically.

## 11. Future Development

- Evaluate event-based triggering or a longer polling interval for CTA Auto-Insert.
- Measure the quality and approval rate of AI-rewritten drafts.
- Track CTA click-through and conversion performance to validate the ranking method.
- Add monitoring for cards that remain in Redo status after repeated failures.
- Review whether the scoring weights should be adjusted based on actual campaign outcomes.
- Add reporting on CTA usage, weekly caps, and the distribution of primary and secondary CTAs.

## 12. Final Outcome

The project establishes a shared content-automation foundation for reviewer-requested rewrites and CTA management.

The live Redo workflow applies feedback to platform-specific drafts, enforces account writing rules, and returns revised content to Airtable for review. The separate CTA Auto-Insert workflow provides a configurable method for ranking and placing CTAs while tracking usage.

---

# Automatic Cleanup of Scheduled Posts When a Card Is Unapproved

> An n8n workflow that checks every 15 minutes for cards that are no longer approved and removes their still-scheduled posts from Metricool, without touching anything already published or about to go out.

Section: n8n Automations

Page: https://mburhan.com/case/n8n-unschedule-cleanup-metricool

## Project Overview

This project automates the cleanup of scheduled social media posts using n8n, Airtable, and Metricool. When a content card is denied, returned to review, or has a scheduling failure, its previously scheduled posts may still be waiting to publish.

The workflow checks Airtable every 15 minutes for cards with statuses such as `Pending Review`, `Denied`, `Schedule Failed`, or a blank status. It retrieves the associated Metricool post IDs, checks each post's current state, and applies safety rules to determine whether it should be deleted, retained, cleared from the card, or checked again later.

The workflow also updates Airtable with the results, clears obsolete post IDs, records timestamped audit entries, and removes the scheduled time when a post is deleted so the content can be scheduled again after approval.

**Technology stack:** n8n, Airtable, Metricool, X, Facebook, LinkedIn, Threads.

## 01. The Business Problem: Unapproved Content Can Still Be Published

A content team manages social media scheduling through Airtable cards and Metricool. Each card stores the Metricool post IDs for its scheduled posts across supported social platforms.

If a card's approval status changes after scheduling, the posts already queued in Metricool may still be published. Without an automated cleanup process, the team must manually locate and unschedule each post.

The main challenges include:

- **Approval changes after scheduling:** Posts can remain queued even after a card is denied or returned for review.
- **Multiple social platforms:** A single card can contain separate Metricool post IDs for X, Facebook, LinkedIn, and Threads.
- **Publishing safety:** Deleting a post that is already published or about to go live can create unnecessary disruption.
- **Repeated processing:** Previously handled post IDs should not be checked repeatedly on every scheduled run.
- **Auditability:** The team needs a record of which posts were deleted, retained, or cleared, and when.
- **Temporary API failures:** A failed lookup or deletion should not prevent the workflow from processing other posts.

The goal was to make scheduled-post cleanup automatic while maintaining strict safeguards around publishing status and timing.

## 02. The Solution: Automated Unscheduling with Safety Checks

A single n8n workflow runs every 15 minutes and searches Airtable for cards that may require cleanup.

The workflow separates each eligible card into individual Metricool post IDs and checks each post sequentially, with a short pause between API calls. A decision step determines whether each post should be deleted, kept, cleared, or retried during a later run.

Only future, unpublished posts that meet the deletion criteria are sent to Metricool's delete endpoint. Published posts and posts due within the next five minutes are protected. Posts that are already overdue but are not marked as published are also left untouched for manual checking.

After processing, n8n updates the corresponding Airtable card, clears IDs for deleted or already-missing posts, records timestamped decisions in the Unschedule Log, and clears the Scheduled Time when a post is deleted.

If a lookup or deletion fails, the workflow leaves the relevant record untouched so that a later run can retry it.

## 03. Project Snapshot

| Item | Detail |
| --- | --- |
| Platform | n8n |
| Schedule | Every 15 minutes |
| Source of approval status | Airtable |
| Social scheduling platform | Metricool |
| Supported networks | X, Facebook, LinkedIn, Threads |
| Input | Airtable cards and Metricool post IDs |
| Decision logic | Status, publishing state, and scheduled time |
| Processing method | One post ID at a time |
| Audit trail | Timestamped Unschedule Log in Airtable |
| Error handling | Failed lookups and deletions remain available for retry |
| Status | Live and running on an automatic schedule |

## 04. How the System Works

### Step 1: Find Cards Requiring Cleanup

Every 15 minutes, the workflow searches Airtable for cards with a status of `Pending Review`, `Denied`, `Schedule Failed`, or blank, provided they still contain a Metricool post ID.

This identifies content that may no longer be eligible for publication but still has scheduled posts associated with it.

### Step 2: Split Cards into Individual Posts

A code node converts each eligible card into one item per Metricool post ID.

The workflow checks its processing log and skips IDs that have already been handled and recorded. This reduces unnecessary repeated processing of settled posts.

### Step 3: Retrieve Each Post from Metricool

The workflow requests the current details of each post from Metricool, processing one post at a time with a short pause between requests.

If an API lookup fails, the error does not stop the entire workflow. The affected post is left untouched so the workflow can try again during a subsequent run.

### Step 4: Apply the Safety Rules

A decision step evaluates the retrieved post information and assigns an action:

- **Delete:** The post is unpublished and scheduled sufficiently far in the future.
- **Keep:** The post is already published or is due within the next five minutes.
- **Clear:** Metricool no longer has the post, so the stale ID can be removed from Airtable.
- **Retry later:** A lookup or deletion fails, leaving the relevant record unchanged for another attempt.

An overdue post that is not marked as published is not automatically deleted. Instead, the workflow records that someone should check its status in Metricool.

### Step 5: Delete Eligible Posts

The `Delete?` branch sends a deletion request to Metricool only for posts marked for deletion.

Posts assigned any other action bypass the deletion request and continue through the merge step. This keeps the delete operation separate from the logic that decides whether deletion is safe.

### Step 6: Update Airtable and Record the Outcome

After the deletion and non-deletion branches merge, a code node prepares one update per Airtable card.

The workflow:

- Clears IDs for posts that were deleted or were already missing from Metricool.
- Adds a timestamped entry to the card's Unschedule Log for each handled post.
- Clears the Scheduled Time when a post is deleted, allowing the content to be scheduled afresh after approval.
- Leaves failed lookup or deletion records unchanged so they can be retried.

Airtable stores the resulting updates, giving the team a record of the cleanup decisions.

## 05. Workflow Architecture

Image (https://mburhan.com/uploads/5a3409d1fd9940b62e685973.png)

The workflow has four main functional layers:

- **Discovery:** Finds Airtable cards that may have obsolete scheduled posts.
- **Post-level processing:** Handles each Metricool post ID separately.
- **Safety and action routing:** Determines whether a post should be deleted, kept, cleared, or retried.
- **Record synchronization:** Updates Airtable IDs, scheduled times, and the audit log.

This separation helps ensure that a post is not deleted simply because its Airtable card has an unsuitable status. Its individual publishing state and scheduled time must also satisfy the safety rules.

### n8n Workflow

Image: The workflow is live and includes protections for published content, posts about to go out, and operations that fail. Its next improvement opportunity is to measure cleanup performance and monitor unresolved cases over time. (https://mburhan.com/uploads/95c97ab4a14597e450c9913b.png)

## 06. Safety Rules and Deletion Logic

The workflow applies the following rules before deleting a scheduled post.

| Post condition | Action | Reason |
| --- | --- | --- |
| Future post, not published, and more than five minutes from its scheduled time | Delete | Eligible for automatic cleanup. |
| Already published | Keep | Published content must not be deleted by this workflow. |
| Scheduled within the next five minutes | Keep | Avoid interfering with content about to go live. |
| Past its scheduled time but not marked as published | Keep and flag for manual checking | Avoid making an unsafe assumption about its status. |
| Metricool no longer has the post | Clear the stale post ID | The post is already gone, so the ID no longer needs to remain on the card. |
| Lookup or deletion request fails | Leave unchanged and retry later | Preserve the record for a subsequent attempt. |

The same safety rules apply across all supported networks, including X.

## 07. Airtable Synchronization and Audit Logging

Airtable remains the central record of approval status and scheduled-post identifiers.

After processing, the workflow synchronizes the card with the actual cleanup outcome.

| Airtable field or record | Purpose |
| --- | --- |
| Card status | Identifies content that may require unscheduling. |
| Metricool post IDs | Tracks the scheduled posts associated with the card. |
| Unschedule Log | Records timestamped actions and decisions. |
| Scheduled Time | Cleared when a post is deleted so it can be scheduled afresh. |

The log provides visibility into what was deleted, retained, or cleared, along with the time of the action. Failed operations remain available for retry rather than being treated as successful cleanup.

## 08. Reliability and Error Handling

The workflow is designed to continue processing other posts when an individual operation encounters an error.

### Sequential API Requests

Metricool is called one post at a time, with a short pause between requests. This reduces the risk of overwhelming the service when many cards require cleanup.

### Failure Isolation

A failed lookup or deletion does not stop the entire workflow. The affected post is left untouched, allowing a later run to retry it.

### Previously Handled IDs

IDs that have already been handled and logged are skipped, reducing repeated checks of settled posts.

### Separate Decision and Deletion Paths

The workflow sends only posts explicitly marked for deletion to Metricool's delete operation. All other decisions bypass the delete request and continue through the merge path.

These controls reduce unnecessary API calls and help protect posts that should remain scheduled. They do not imply that API failures or every possible scheduling edge case are eliminated.

## 09. Results and Operational Capabilities

The project description confirms that the workflow is live and runs automatically. It does not provide measured figures for the number of posts cleaned up, API success rate, time saved, or prevented accidental publications.

### Implemented Capabilities

- Automatic checks every 15 minutes.
- Identification of Airtable cards requiring scheduling cleanup.
- Individual processing of Metricool post IDs.
- Skipping of previously handled and logged IDs.
- Protection for published posts and posts due within five minutes.
- Manual-review handling for overdue posts not marked as published.
- Removal of eligible future, unpublished posts.
- Cleanup of stale post IDs when posts are already missing.
- Timestamped Unschedule Log entries.
- Scheduled Time clearing after deletion.
- Retry behavior for failed lookups and deletion requests.

### Metrics to Measure Next

A measured production review could track:

- Number of cards and posts evaluated per run.
- Number of posts deleted, retained, cleared, and retried.
- Average time between an approval-status change and cleanup.
- Number of failed Metricool requests.
- Number of posts that require manual status checks.
- Number of duplicate or unnecessary cleanup attempts.
- Any cases where an unapproved post remained scheduled or a valid post was incorrectly targeted.

These figures would help quantify the operational value and verify the effectiveness of the safety rules.

## 10. Key Takeaways

- **Approval status must remain synchronized with scheduling:** Changing an Airtable card does not automatically remove a post already queued in Metricool.
- **Evaluate individual posts before deletion:** A card's status alone is not enough to determine whether every associated post can safely be removed.
- **Protect publishing boundaries:** Published posts, posts due within five minutes, and ambiguous overdue posts receive special handling.
- **Keep an audit trail:** Timestamped Unschedule Log entries make cleanup decisions visible to the team.
- **Handle failures without losing state:** Failed operations remain available for a later retry.
- **Use sequential API calls:** Processing posts one at a time helps control the request rate.

## 11. Future Development

The next stage should focus on measuring reliability and improving operational visibility.

- Add a periodic summary of deleted, retained, cleared, and retried posts.
- Track repeated failures and alert the team when a post remains unresolved across multiple runs.
- Measure the time between card-status changes and successful cleanup.
- Review manual-check cases to determine whether additional Metricool status signals could safely resolve them.
- Test the workflow against edge cases involving multiple networks, missing post IDs, schedule changes, and posts close to their publication time.
- Maintain an audit of cleanup outcomes to support ongoing verification of the deletion safeguards.

## 12. Final Outcome

The Automatic Cleanup of Scheduled Posts workflow provides an automated safeguard between Airtable approval decisions and Metricool's publishing queue.

By checking cards every 15 minutes, evaluating each associated post, applying time- and status-based deletion rules, and recording the outcome in Airtable, it helps prevent obsolete scheduled content from being left in the publishing queue.

---

# Telegram Thread Builder: From Dropped Screenshots to a Reviewed X Thread

> An n8n workflow that turns a source text and screenshots dropped in a Telegram chat into a ready-to-review X thread, with automatic image blurring, AI writing and a card created in Airtable.

Section: n8n Automations

Page: https://mburhan.com/case/n8n-telegram-thread-builder

## Telegram Thread Builder: From Dropped Screenshots to a Reviewed X Thread

**Transform source text, screenshots, and team notes shared in Telegram into AI-generated X threads or single posts, with automatic image blurring and an Airtable review workflow.**

## Project Overview

The Telegram Thread Builder is an n8n automation designed for content teams that collect ideas, source material, and screenshots in a Telegram group. It turns these inputs into structured X threads or individual posts without requiring the team to leave Telegram during content preparation.

The workflow manages the entire process, from collecting source material and asking setup questions to blurring images, generating content with Claude, applying calls to action, and creating a review card in Airtable.

Each session is tracked in Airtable, allowing the workflow to preserve messages, images, captions, and notes throughout the process. The resulting content is marked **Pending Review**, giving the team an opportunity to inspect the draft and verify image blurring before approval.

**Technology stack:** n8n, Telegram, Airtable, Claude, image-blurring service, X content workflow.

## 01. The Business Problem: Turning Unstructured Ideas into Publishable Content

Content teams often collect information in chat conversations, with source text, screenshots, captions, and notes spread across multiple messages. Turning this material into a coherent X thread requires manual organization, writing, image preparation, formatting, and review.

The process creates several challenges:

- **Scattered source material:** Text, screenshots, and captions may arrive in separate Telegram messages.
- **Manual content creation:** Writers must organize the material into a logical sequence and maintain a consistent voice.
- **Image privacy:** Screenshots may contain page names, logos, or faces that need to be blurred before use.
- **Inconsistent formatting:** Threads must follow an agreed post count, image order, writing style, and call-to-action placement.
- **Lost context:** Without session tracking, it becomes difficult to preserve all the material collected during a conversation.
- **Review overhead:** Drafts, images, and approvals need to be organized in a central location.

The goal was to turn this unstructured collection process into a guided, repeatable workflow with safeguards and a clear human-review stage.

## 02. The Solution: Telegram-Based Content Generation and Review

The workflow uses Telegram as the team's conversational interface. A user starts a session, submits text and screenshots, answers a small set of configuration questions, and triggers the content-generation process.

n8n stores session materials in Airtable, retrieves the images in order, sends them through an image-blurring service, and verifies that the required blurring has succeeded. Claude then receives the source material, blurred images, captions, notes, and current writing rules from Airtable.

After generation, code nodes clean and validate the draft, enforce the requested post count, check image order, and insert calls to action from the Airtable CTA library.

The workflow creates an Airtable review card containing the draft, hook, and blurred images. Telegram then returns a summary with a preview and a link to the card.

The system also supports single-post requests by routing them through the team's existing content-intake workflow.

## 03. Project Snapshot

| Item | Detail |
| --- | --- |
| Platform | n8n |
| Input channel | Telegram group chat |
| Content destination | X |
| AI writing | Claude |
| Session and content storage | Airtable |
| Image processing | External image-blurring service |
| Workflow size | Approximately 65 nodes |
| Session duration | 12 hours |
| Image capacity | Up to 15 images per session |
| Setup timeout | 10 minutes before defaults are applied |
| Output formats | Multi-post X thread or single post |
| Review status | Pending Review |
| Rule and CTA management | Airtable |

## 04. How the System Works

The workflow follows five main stages, from collecting the source material to delivering a review-ready draft.

### Step 1: Start a Session and Collect Materials

A team member replies `Go` to a submitted text or image. The message can also specify a format, such as `thread`, `image`, `text`, or `long`.

The bot explains how to provide the source material:

- Send screenshots in the required order.
- Add captions beneath images when a particular caption should become the text for that post.
- Send additional source text or notes.
- Send `done` when the material is ready.
- Send `cancel` to stop the session.

Telegram may split long text into multiple messages. The workflow merges those parts back together in the correct order.

Each image and text item is saved as a separate Airtable row linked to the session. Sessions remain open for 12 hours.

### Step 2: Configure the Content

Once the user sends `done`, the bot asks only the questions that remain unanswered.

These can include:

- How many posts should the thread contain?
- Which call to action should be used, and where should it appear?
- Should the images be blurred?

The questions are numbered so the user can respond with a simple number. Typing `go` accepts the defaults for any remaining questions, while options supplied with the initial `Go` command can skip questions entirely.

The bot summarizes the selected configuration. Replying `1` starts the build.

If the user does not respond within 10 minutes, the workflow proceeds automatically using the defaults.

### Step 3: Prepare and Blur Images

The workflow retrieves the session's images in their original order, up to a maximum of 15, and downloads them from Telegram.

Each image is sent to an external blurring service designed to hide page names, logos, and faces.

The workflow then checks whether every image has returned successfully blurred. If any image fails, content generation is stopped, the session is reopened, and Telegram identifies the failed images and reasons where available.

The user can correct the issue and send `done` again rather than restarting the entire session.

If the team explicitly chooses `no blur`, the workflow follows that configuration instead.

### Step 4: Generate the Thread with Claude

Claude receives the source text, team captions and notes, blurred images, and current Post Rules from Airtable.

The writing instructions specify the requested number of posts, preserve the intended image order, and require the writer to use the facts contained in the supplied material, including relevant numbers and labels in screenshots.

The requested writing style includes:

- A strong, extended opening hook.
- Follow-up posts that each contribute one step to the story.
- An analytical post near the end.
- Consistent use of the agreed content format.
- No naming of pages, people, or brands shown in the source material.

These constraints help turn collected material into a coherent thread rather than a sequence of unrelated summaries.

### Step 5: Validate, Add CTAs, and Deliver

After Claude generates the content, a code node cleans the output by removing unwanted links, handles, attribution credits, post numbering, and stock phrases.

It then checks the post count and verifies that each image is used exactly once and in the correct order. If the image sequence is incorrect, the workflow redistributes the images.

Calls to action are selected from the Airtable CTA library. Examples include a newsletter CTA at the end of a thread or a consulting CTA placed at post four.

The workflow creates an Airtable card containing:

- The opening hook.
- The complete thread or post.
- The selected CTA.
- The associated blurred images.
- The status `Pending Review`.

Finally, the session is marked complete and Telegram sends a summary containing the post count, image count, selected CTA, a preview of the hook, and a link to the review card.

## 05. Workflow Architecture

Image (https://mburhan.com/uploads/1ee61b70612225e17e1aa5de.png)

The workflow contains approximately 65 nodes, organized into functional sections.

- **Input and session management:** Interprets Telegram messages and manages session state.
- **Material collection:** Stores text, captions, notes, and images in Airtable.
- **Configuration:** Resolves post count, CTA placement, and image-blurring preferences.
- **Image processing:** Downloads, blurs, and validates screenshots.
- **AI writing:** Sends source material and current writing rules to Claude.
- **Content validation:** Cleans the output, checks post count, and verifies image usage.
- **Delivery and review:** Creates the Airtable card, attaches images, and returns the card link through Telegram.

Additional branches manage configuration timeouts, cancellation, retry handling, and single-post requests.

Image: The Telegram thread session workflow in n8n: messages come in on the left and are routed by what they mean (start, image, note, cancel, done, ask, single post). The long middle row builds the thread, from blurring the images to the Claude writer, the Airtable card and the Telegram summary. The rows below handle the setup questions with the 10-minute timer, and single-post sessions. (https://mburhan.com/uploads/3d24cd76c96ab74d1df41d76.png)

## 06. Single-Post Generation

The workflow also supports individual posts that do not require a multi-post thread.

After the user sends `Go`, the bot asks the shorter set of configuration questions. The workflow combines the source text, captions, first image, and selected CTA into a single request.

That request is passed to the team's existing content-intake workflow, allowing single posts to reuse the established writing, image-blurring, and Airtable-card creation steps.

This keeps the single-post path aligned with the thread-generation process without duplicating the entire content pipeline.

## 07. Built-In Safeguards and Validation

The workflow includes several controls designed to prevent incomplete or unusable content from reaching the review stage.

### Image-Blurring Validation

- Images are processed before they are attached to the generated content.
- The workflow checks the returned images for successful blurring.
- If a required blur operation fails, thread generation stops.
- The session is reopened so the team can retry after addressing the issue.

The team can explicitly choose not to blur images where appropriate.

### AI Output Validation

- The generated content is cleaned before delivery.
- The final post count is checked against the agreed configuration.
- Images must be used exactly once and in the correct order.
- If the writer returns unusable output, no review card is created.

### Failure Recovery

- When blurring or writing fails, the session is reopened and the team is notified through Telegram.
- The user can retry using `done` without starting a new session.
- Image-upload failures are reported in the final summary.

These safeguards are designed to prevent incomplete drafts from being treated as ready for review. They do not replace a human check of image privacy, factual accuracy, and content quality.

## 08. Airtable as the Content Control Center

Airtable stores the information needed to manage sessions and control the writing process.

| Airtable component | Purpose |
| --- | --- |
| Session records | Track the active content-building session. |
| Material rows | Store submitted text, captions, notes, and images against the session. |
| Post Rules | Provide the current writing instructions used by Claude. |
| CTA library | Supply approved calls to action and placement options. |
| Review cards | Store the generated hook, thread, images, and review status. |

Because Post Rules, CTAs, and session materials are stored in Airtable, the team can update writing preferences and content options without changing the n8n workflow itself.

## 09. Results and Deliverables

The supplied project description documents the workflow's design and capabilities but does not include measured production metrics such as hours saved, generation success rate, content approval rate, or publishing performance.

### Delivered Capabilities

- Telegram-based session creation and content collection.
- Support for text, screenshots, captions, and notes.
- Automatic merging of split Telegram text messages.
- Session persistence through Airtable.
- Configuration questions with default values and timeout handling.
- Image blurring and validation before content generation.
- Claude-powered thread and single-post generation.
- Automated post-count and image-order checks.
- CTA selection from an Airtable library.
- Airtable review-card creation with `Pending Review` status.
- Telegram delivery of a summary and review link.
- Retry and recovery paths for several processing failures.

### Metrics to Track Next

To quantify the business impact, a pilot should measure:

- Average time required to produce a thread manually versus using the workflow.
- Number of sessions and drafts generated per week.
- Percentage of sessions completed successfully.
- Image-blurring success and retry rates.
- Percentage of generated drafts approved without major edits.
- Average number of content revisions per draft.
- Time from initial Telegram submission to review-card creation.

These measurements would help establish whether the automation reduces preparation time while maintaining the team's content and privacy standards.

## 10. Key Takeaways

- **Use chat as the intake interface:** Teams can submit ideas and screenshots where they already collaborate.
- **Preserve session context:** Saving every item in Airtable reduces the risk of losing source material across multiple messages.
- **Automate repetitive preparation:** Image handling, content formatting, CTA insertion, and card creation happen within the same workflow.
- **Validate before delivery:** Image checks and output validation reduce the risk of incomplete drafts reaching review.
- **Keep writing rules configurable:** Airtable allows the team to update Post Rules and CTAs without editing the automation.
- **Retain human review:** The workflow prepares content for approval instead of automatically publishing it to X.

The next stage should focus on measuring reliability, improving review visibility, and refining content quality.

## 11. Future Development

- Track generation success, retry rates, and average processing time.
- Record review outcomes and common reasons for content revisions.
- Add clear audit information for image-blurring failures and content-validation failures.
- Improve visibility into session expiration and abandoned sessions.
- Evaluate whether generated threads consistently follow the team's style and factual requirements.
- Consider adding a publishing integration only if the team wants to extend the current review-first process.

## 12. Final Outcome

The Telegram Thread Builder converts source text, screenshots, and team notes into structured X threads or single posts through a guided conversational workflow.

By combining n8n, Telegram, Claude, and Airtable, it automates session management, image preparation, AI writing, content validation, CTA insertion, and review-card creation.

The result is a repeatable content-preparation process that keeps source material organized, applies image-blurring safeguards, and gives the team a central place to review generated drafts before publication.

---

# Newsletter Performance Agent: Daily Scoring of Every Sent Issue

> An n8n workflow that finds each newsletter issue after it is sent, removes machine clicks from the numbers, scores it against the account’s own baseline, and posts an AI analysis to the team.

Section: n8n Automations

Page: https://mburhan.com/case/n8n-newsletter-performance-agent

## Newsletter Performance Agent: Daily Scoring of Every Sent Issue

**Automatically analyze newsletter performance, filter machine-generated clicks, compare each issue against historical benchmarks, and deliver AI-powered insights to the team.**

## Project Overview

The Newsletter Performance Agent is an automated analytics workflow built with n8n for a publisher using Kit to distribute regular email newsletters. It identifies newly sent newsletter issues, retrieves their performance statistics, adjusts click metrics for suspected machine traffic, and compares each issue against the account's historical performance.

Claude analyzes the results using the account's existing newsletter rules and returns a structured performance assessment with evidence-based recommendations. The workflow stores the complete analysis in Airtable, sends a concise summary to the team through Telegram, and records each execution for tracking.

The system is designed to support better newsletter decisions without allowing AI-generated recommendations to become active rules automatically.

**Technology stack:** n8n, Kit, Airtable, Claude, Telegram.

## 01. The Business Problem: Reliable Newsletter Performance Analysis

Newsletter performance data can be misleading. Email security gateways may automatically click links to inspect their safety, inflating reported engagement and making it difficult to understand what readers actually clicked.

Comparing issues also requires a consistent baseline. A newsletter sent to a small segment should not necessarily be judged against a full-list send, and raw engagement figures alone cannot explain whether a particular issue performed better or worse than usual.

The main challenges were:

- **Machine-generated clicks:** Automated link scanning can inflate click-through rates.
- **Inconsistent comparisons:** Each newsletter needs to be evaluated against the account's own historical performance.
- **Limited link-level insights:** Product interest and link placement need to be analyzed across campaigns.
- **Repeated manual reporting:** Statistics, comparisons, and recommendations otherwise require separate steps.
- **Risk of unsupported recommendations:** AI must not invent data, treat clicks as sales, or recommend rules already in force.

## 02. The Solution: Automated Performance Scoring and AI Analysis

A single n8n workflow runs daily at 17:00 South African time and can also be started manually.

The workflow retrieves recent issue records from Airtable and broadcasts from Kit, selects the newest completed send from the last 36 hours that has not already been logged, and collects its performance and link-click data.

Code nodes calculate adjusted metrics, establish historical comparisons, attribute clicks to products and positions, and flag uncertain data. Claude then analyzes the prepared results against the active newsletter rules stored in Airtable.

The final report is saved to Airtable. A short summary is sent to the team's Telegram feedback group, each run is logged, and any proposed rules are saved as inactive records for human review.

## 03. Project Snapshot

| Item | Detail |
| --- | --- |
| Platform | n8n |
| Newsletter service | Kit |
| Analytics and storage | Airtable |
| AI analysis | Claude |
| Team notifications | Telegram |
| Schedule | Daily at 17:00 South African time, with manual execution available |
| Issue selection | Newest completed send within the previous 36 hours that has not been logged |
| Historical context | Last 40 logged issues in Airtable |
| Click analysis | Machine-click adjustment and link-level attribution |
| Output | Performance score, AI analysis, recommendations, and run log |
| Rule activation | Manual approval required |

## 04. How the System Works

The workflow follows a series of stages from issue selection through performance analysis and reporting.

### Step 1: Scheduled Execution and Historical Data

The workflow starts on its daily schedule or through a manual trigger. It reads the last 40 logged issues from Airtable and retrieves the available broadcasts from Kit.

### Step 2: Identify the Newsletter Issue

The workflow selects the newest completed send within the previous 36 hours that has not already been recorded.

If no eligible issue exists, the workflow stops without producing a score. The Kit broadcast ID and send date are used to prevent an issue from being scored more than once.

### Step 3: Retrieve Performance Statistics

The workflow retrieves the selected issue's performance statistics and link-level click report from Kit.

If link-level data is unavailable, the workflow continues with raw click figures and marks them as unverified. This uncertainty is carried into the subsequent analysis.

### Step 4: Calculate and Adjust Metrics

A code node removes suspected machine-generated clicks, calculates performance against the account's historical baseline, and attributes link clicks to product campaigns and newsletter positions.

The calculation also identifies conditions that could affect interpretation, including segment sends and missing campaign tags.

### Step 5: Generate the AI Analysis

The workflow retrieves active newsletter rules from Airtable and sends Claude the calculated metrics, relevant historical context, and existing rules.

Claude returns structured analysis. A subsequent code node cleans the output, limits proposed rules, and labels each recommendation as `PROPOSAL, not active`.

### Step 6: Save and Share the Results

The completed performance report is saved as a new record in Airtable's Performance table, including statistics, score, analysis, and recommendations.

Two follow-up branches then run:

- **Team reporting:** Sends a concise summary to the Telegram feedback group and records the workflow run.
- **Rule management:** Stores proposed rules in Airtable's Newsletter Rules table as inactive entries.

## 05. Workflow Architecture

Image (https://mburhan.com/uploads/f4d9bbd87ecc13bb3363e2af.png)

The workflow separates deterministic calculations from AI interpretation. Code handles the numerical metrics and data-quality flags, while Claude explains the results and proposes potential improvements.

Airtable serves as the historical record and source of active newsletter rules. Kit provides newsletter and performance data, and Telegram provides the team's concise reporting channel.

### n8n work Flow:

Image (https://mburhan.com/uploads/f8811394ebb0ed5eb92bdb4d.png)

## 06. Performance Scoring and Click Analysis

### Machine-Click Adjustment

Email security gateways may automatically click links in a newsletter. This can create a flat plateau across multiple links, unlike the steeper click distribution that may be associated with genuine reader activity.

The workflow detects this plateau and subtracts the suspected machine traffic to calculate an adjusted click rate.

The supplied examples demonstrate the effect:

| Example | Raw click rate | Adjusted click rate |
| --- | --- | --- |
| Historical issue example 1 | 7.73% | 2.63% |
| Historical issue example 2 | 6.64% | 2.54% |
| Clean issue | No specific figure supplied | Left unchanged |

These figures illustrate the adjustment logic. They do not establish an independently verified measure of human clicks; the adjusted figures remain dependent on the workflow's machine-traffic detection method.

### Historical Baseline Scoring

Each issue is compared against the median open rate and click rate of past issues.

Segment sends with fewer than half the usual recipients are excluded from the baseline comparison. This helps prevent smaller, targeted sends from distorting comparisons with normal full-list newsletters.

A score of `1.00` represents performance at the established baseline.

### Product and Position Attribution

The workflow groups clicks using two dimensions:

- **Campaign tag:** Identifies the associated product or campaign, with the date suffix removed so results can be compared over time.
- **Slot tag:** Identifies the position where a link appeared in the newsletter.

This structure supports analysis of which product campaigns attract interest and how newsletter placement relates to engagement.

### Data-Quality Flags

The workflow flags conditions that may limit confidence in the analysis, including:

- Missing link-level click data.
- Segment sends.
- Links without a dated campaign tag.

These flags are passed to Claude so uncertainty can be reflected in the final report rather than hidden.

## 07. AI Analysis and Guardrails

Claude receives the calculated performance data and active newsletter rules from Airtable. Its task is to explain the results and identify potential improvements without overstating what the data supports.

### Analysis Rules

- **Use adjusted clicks as the primary measure:** Raw clicks must not be presented as a performance win when they may contain machine traffic.
- **Explain discrepancies:** If raw and adjusted click rates tell different stories, the analysis must acknowledge the difference.
- **Do not equate clicks with revenue:** Sales cannot be measured by this workflow, so product-link clicks represent interest, not confirmed sales.
- **Avoid duplicate recommendations:** Existing active rules must not be proposed again.
- **Require evidence:** Each proposed rule must state the evidence supporting it.
- **Allow no recommendation:** Zero proposed rules is an acceptable outcome when the evidence is insufficient.
- **Keep reports concise:** The analysis is limited to two paragraphs, with no more than three recommendations. The Telegram summary must remain under 600 characters.
- **Do not invent data:** Claims must be grounded in the supplied statistics and context.

### Human Approval Before Activation

The agent recommends; it does not make policy decisions.

Proposed rules are stored as inactive records in Airtable. A person must explicitly mark a proposal as Active before it can reach the newsletter writing prompts.

This keeps the workflow useful for analysis while preserving human control over changes to newsletter strategy.

## 08. Reporting and Data Management

The workflow creates a consistent reporting process across three connected outputs.

| Component | Purpose |
| --- | --- |
| Airtable Performance table | Stores issue statistics, score, analysis, and recommendations. |
| Telegram feedback group | Receives a short summary for the team. |
| Execution log | Records workflow runs for operational tracking. |
| Airtable Newsletter Rules table | Stores proposed rules as inactive until approved. |

The Kit broadcast ID and send date provide the issue-level identifiers used to prevent repeat scoring. The last 40 logged issues supply recent historical context, while active rules provide the current policy context for Claude.

## 09. Results and Metrics

The supplied workflow description documents the automation's capabilities and includes examples of machine-click adjustment. It does not provide a measured before-and-after evaluation of time saved, newsletter engagement improvement, recommendation adoption, or revenue impact.

### Demonstrated Workflow Capabilities

- Daily and manual execution.
- Selection of eligible, previously unlogged newsletter issues.
- Retrieval of Kit statistics and link-level click data.
- Machine-click adjustment and historical baseline scoring.
- Product and position-level click attribution.
- Data-quality flags passed into the AI analysis.
- Structured Claude output and bounded recommendations.
- Airtable reporting, Telegram summaries, and execution logging.
- Human approval before proposed rules become active.

### Example Metric Adjustment

The provided examples show raw click rates of 7.73% and 6.64% being adjusted to 2.63% and 2.54%, respectively, while a clean issue is left unchanged.

These are examples of the calculation behavior, not evidence of improved campaign performance.

### Metrics to Measure Next

To evaluate business impact, a pilot should record:

- Time spent producing newsletter reports before and after automation.
- Percentage of eligible issues successfully scored.
- Frequency of missing or unverified link reports.
- Agreement between adjusted click estimates and a trusted validation sample.
- Frequency and quality of AI recommendations.
- Percentage of proposals approved by a human.
- Changes in engagement across subsequent issues, evaluated separately from the workflow's own scoring.

## 10. Key Takeaways

- **Correct the data before interpreting it:** Machine-generated clicks can make newsletter performance appear stronger than it is.
- **Use an account-specific baseline:** Comparing each issue with historical medians provides context beyond raw statistics.
- **Keep calculations separate from AI analysis:** Deterministic code handles metrics, while Claude interprets the prepared evidence.
- **Carry uncertainty forward:** Missing reports and questionable attribution are flagged rather than silently treated as verified.
- **Keep recommendations evidence-based:** The agent must explain its proposals and can return no recommendations when evidence is insufficient.
- **Preserve human control:** Proposed rules remain inactive until explicitly approved.
- **Make reporting repeatable:** Airtable, Telegram, and execution logs provide structured records for the team.

## 11. Future Development

The next stage should focus on validating the scoring method and measuring its operational value.

- Compare adjusted click rates with a manually reviewed sample to evaluate the machine-click detection method.
- Track report completion and data availability across eligible newsletter issues.
- Measure the time required to prepare reports manually versus using the workflow.
- Track which recommendations are accepted, rejected, or modified by the team.
- Evaluate subsequent newsletter engagement to determine whether approved recommendations are associated with meaningful improvements.
- Monitor cases where link-level data is missing or campaign and slot tags are incomplete.

## 12. Final Outcome

The Newsletter Performance Agent provides an automated process for selecting newly sent newsletter issues, calculating adjusted engagement metrics, comparing performance with historical benchmarks, and generating a concise AI analysis for the team.

By combining Kit, n8n, Airtable, Claude, and Telegram, the workflow turns newsletter reporting into a repeatable process with explicit data-quality flags, structured recommendations, and human approval controls.

The key distinction is that the system automates analysis and recommendation management; improvements in engagement, efficiency, and revenue still need to be validated through measured use.

---

# WhatsApp Voice & Business Agent with n8n

> An AI-powered WhatsApp assistant that understands voice notes, text messages, and photos, then answers customer questions using product information stored in Supabase.

Section: n8n Automations, AI Agents for WhatsApp

Page: https://mburhan.com/case/whatsapp-voice-business-agent

## Overview

This project is a WhatsApp business assistant for an online store. Customers can send a voice note, text message, or product photo and receive an answer based on the store’s product database. The system uses n8n to process incoming messages, OpenAI to understand different input types, and Supabase to retrieve product information. When a customer sends a voice note, the assistant can respond with a voice message. Conversation memory also helps maintain context across messages from the same customer.

### The business problem

Online stores receive repeated questions about product availability, prices, and delivery. Customers may ask these questions through text, voice notes, or photos, making them difficult for simple chatbots to handle. Manual replies take time, while inaccurate answers about prices or stock can reduce customer trust.

## How It Works

1. **Receive WhatsApp Messages:** The WhatsApp Business Cloud API receives customer messages, including text, voice notes, and photos.
2. **Identify Message Type:** n8n checks the incoming message and routes it to the correct processing path.
3. **Process Voice and Images:** OpenAI transcribes voice notes and analyzes product photos. Text messages are passed through directly.
4. **Prepare the Message:** n8n converts each input into a common text format and associates it with the customer's phone number.
5. **Retrieve Product Information:** The AI agent uses the Supabase product database to find relevant information before preparing an answer. Conversation memory helps it maintain context.
6. **Send the Response:** Text messages receive text replies. When the customer sends a voice note, the answer is converted into speech and returned as a WhatsApp voice message.

Image (https://mburhan.com/uploads/30dbed1dba656a505c6917ae.png)

## The Result

The completed workflow brings voice, text, and image-based customer enquiries into one WhatsApp assistant. It retrieves product information from Supabase before answering, supports both text and voice replies, and maintains conversation context for each customer.

### Workflow Delivery Metrics

| Metric | Result |
| --- | --- |
| n8n nodes after cleanup | 21 |
| Supported input types | 3 |
| Reply formats | 2 |
| AI models used | 4 |
| Main agent workflows | 1 |
| Workflow defects fixed | 2 |

*Note: These figures describe the delivered workflow. Live customer response time, answer accuracy, and sales conversion have not yet been measured.*

Image (https://mburhan.com/uploads/272f0e7121790fb2c057e5e0.png)

## Tech Stack

- **n8n:** Connects the services and manages the message-processing workflow.
- **WhatsApp Business Cloud API:** Receives customer messages, downloads media, and sends replies.
- **OpenAI:** Transcribes voice notes, analyzes images, generates answers, and converts replies into speech.
- **Supabase:** Stores product information and provides a searchable knowledge source for the AI agent.
- **Shopify:** Relevant to the planned extension for product and order-related functionality; order lookup and replacement-order creation are not part of the current workflow.

## Where This Pattern Applies

- **E-commerce Stores:** Answer product questions about prices, availability, and product details.
- **Retail Businesses:** Help customers identify products through photos and ask questions through voice or text.
- **Customer Support Teams:** Handle common enquiries and pass complex issues to human support.
- **Online Service Providers:** Provide customers with a WhatsApp-based assistant for frequently asked questions.

This pattern is useful for businesses that receive customer enquiries through WhatsApp and want to automate responses while keeping answers connected to reliable business data.

---

# Support Email to ClickUp Ticket Automation

> An automated support workflow that turns incoming customer emails into correctly routed ClickUp tickets, adds replies to existing tickets, and prevents duplicate tickets using n8n.

Section: n8n Automations, AI Agents for Ecommerce, ClickUp Setups

Page: https://mburhan.com/case/support-email-clickup-automation

## Overview

This solution automates support email handling for RefureTech. When a customer sends an email to the shared Gmail inbox, n8n reads the message, identifies the client using their email domain, and sends it to the correct ClickUp support list. If the client is unknown, the email goes to a default triage list. If the conversation already has a ticket, the new email is added as a comment instead. Attachments and previous messages are also included, giving the support team the full conversation in one place.

### The business problem

Support requests arrived through a shared email inbox and had to be added to ClickUp manually. This took time and created a risk of assigning emails to the wrong client, missing attachments, or creating duplicate tickets. ClickUp also has its own email feature that handles customer replies, so a basic email-to-ticket automation could create duplicate tickets or process its own confirmation emails repeatedly.

### The solution

A single n8n workflow polls the support inbox every minute. It looks up the client list stored in ClickUp, matches the sender by domain, checks four different ways whether the conversation already has a ticket or is already managed by ClickUp, and only then creates a ticket or comment. It uses plain, explainable rules in code nodes rather than an AI model, so every decision can be traced to a header, a subject line or a ClickUp field. Wherever a step could fail, the workflow prefers a duplicate ticket over a lost customer email.

### How It Works

1. **Receive Support Emails:** The Gmail Trigger checks the shared inbox every minute and retrieves new messages, including their attachments.
2. **Identify the Client:** n8n checks the sender's email domain against the client information stored in ClickUp. Each client has a domain, folder ID, and support list ID.
3. **Check for Existing Tickets:** The workflow searches for existing tickets using Gmail thread IDs and ClickUp email thread IDs. It also checks the full email conversation to identify related messages.
4. **Decide Where the Email Goes:** The workflow uses routing rules to determine whether the email is a new request, a reply to an existing ticket, an unknown client request, or a message that should be skipped.
5. **Create a Ticket or Add a Comment:** New requests are added to the correct client's support list. Emails from unknown clients go to the default triage list, while replies to existing tickets are added as comments.
6. **Upload Attachments:** Email attachments are downloaded and uploaded to the related ClickUp ticket. Earlier messages in the conversation are also included where available.
7. **Prevent Duplicate Processing:** The workflow skips ClickUp confirmation emails, messages already managed by ClickUp, and automatic replies such as out-of-office messages.
8. **Mark Emails as Processed:** Successfully handled emails receive a `Ticketed` label in Gmail. For new tickets, the workflow also waits two minutes and saves ClickUp's confirmation email thread ID for future matching.

Image (https://mburhan.com/uploads/8e55d3a30801ccfad023f7ba.png)

### The Result

The automation successfully manages incoming support emails, creates and routes ClickUp tickets, and adds replies to existing tickets. It also preserves attachments and email history while preventing duplicate tickets and unnecessary processing.

**Automation Performance**

| Metric | Result |
| --- | --- |
| Successful workflow executions | 11 |
| Successful execution rate | 100% |
| Failed executions | 0 |
| Total n8n nodes | 38 |

### Tech Stack

- **n8n:** Manages the automation, email processing, routing rules, and duplicate checks.
- **Gmail:** Receives support emails, retrieves message threads and attachments, and labels processed messages.
- **ClickUp:** Stores client information, creates support tickets, adds comments, and manages attachments.
- **JavaScript:** Handles client matching, email validation, routing decisions, and duplicate checks.
- **ClickUp API:** Supports searches and operations that connect the workflow with ClickUp records.

### Where This Pattern Applies

The same approach can be used by organizations that manage customer requests through shared email inboxes.

- **IT Support Teams:** Convert incoming technical support emails into tickets.
- **Customer Service Teams:** Route customer requests to the correct team or department.
- **Managed Service Providers:** Organize support requests across multiple clients.
- **SaaS Companies:** Track customer issues and follow-up conversations in one system.
- **Operations Teams:** Turn shared inbox requests into assigned and trackable tasks.

The general approach is to monitor a shared inbox, identify the customer, check for existing conversations, and automatically create or update the right ticket while keeping the full email history and attachments.

---

# HubSpot Industry Classifier with Gemini

> An automated workflow that classifies HubSpot companies into industry groups using Google Gemini and sends uncertain records for human review.

Section: n8n Automations, AI Agents

Page: https://mburhan.com/case/hubspot-industry-classifier

## Overview

This project uses n8n and Google Gemini Flash to automatically classify company records in HubSpot. The workflow runs every hour, retrieves companies from a designated HubSpot list, and checks whether their industry group is already assigned. Gemini analyzes company information and suggests an industry based on the categories available in HubSpot.

To maintain CRM data quality, the workflow validates every suggested category and checks its confidence score before updating a record. Companies that cannot be classified reliably are marked as **Needs R&D** for human review. A dry-run mode also allows the team to test the workflow without changing live CRM records.

## The Problem

Many HubSpot company records may have missing or inconsistent industry classifications. Updating these records manually takes time and can lead to inconsistent categories.

Using AI without proper checks also creates risks, such as invented industry groups, unreliable classifications, and repeated processing of companies that have already been classified.

## How It Works

1. **Scheduled Execution:** n8n starts the workflow every hour and loads its configuration, including the company limit, confidence threshold, and dry-run setting.
2. **Retrieve Company Records:** The workflow retrieves companies from a designated HubSpot list, collecting up to 100 records per run.
3. **Check Company Information:** It retrieves each company's name, domain, website, and existing industry group. Already classified companies and records without usable information are skipped.
4. **AI Classification:** Google Gemini Flash analyzes the available company information and returns a suggested industry group, confidence score, exclusion flag, and reasoning.
5. **Validate the Result:** n8n checks whether the suggested industry exists in HubSpot's allowed options. It also applies a minimum confidence threshold of 0.6 and limits confidence to 0.65 for companies without a domain.
6. **Update or Flag Records:** Valid classifications are written to HubSpot with relevant classification details. Uncertain or excluded companies are marked as Needs R&D for human review.
7. **Control Updates:** Dry-run mode prevents changes to live records during testing. A Wait node also separates updates to help manage HubSpot API rate limits.

Image (https://mburhan.com/uploads/480b3d2db06d67347fe2909b.png)

## The Result

The workflow provides an automated and controlled process for classifying HubSpot company records. It uses Gemini to suggest categories while HubSpot provides the approved industry options. Validation rules, confidence checks, and human-review routing help prevent unreliable classifications from being accepted automatically.

### Workflow Configuration Metrics

| Metric | Result |
| --- | --- |
| n8n workflow nodes | 22 |
| Execution schedule | Every hour |
| Maximum companies per run | 100 |
| Confidence threshold | 0.6 |
| No-domain confidence cap | 0.65 |
| AI model | Google Gemini Flash |

*Note: These figures describe the workflow configuration. Actual companies processed, classification accuracy, human-review rate, and time saved have not yet been measured.*

## Tech Stack

- **n8n:** Runs the scheduled automation and manages the classification process.
- **HubSpot:** Provides company records, approved industry groups, and the destination for validated updates.
- **Google Gemini Flash:** Analyzes company information and suggests industry classifications with confidence scores and reasoning.

## Where This Pattern Applies

- **CRM Management Teams:** Keep company industry fields consistent across large CRM databases.
- **Sales Teams:** Organize company records for better segmentation and reporting.
- **Marketing Teams:** Group companies into relevant industries for targeted campaigns.
- **Business Data Teams:** Automate data enrichment while routing uncertain records for manual review.

This pattern is useful for businesses that need to enrich large datasets with AI while maintaining control over data quality and ensuring uncertain results receive human review.

---

# AI-Powered Fashion Campaign Image Generation

> An n8n workflow that turns one product photo into AI-generated model images plus eCommerce and lifestyle campaign shots, saved to Google Drive.

Section: n8n Automations, AI Agents, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-fashion-campaign-image-generation

## Project Overview

The AI-Powered Fashion Campaign Image Generation system automates fashion visual production using n8n, Google Gemini, image generation APIs, and Google Drive. Starting with a single uploaded clothing image, the workflow analyzes the garment, develops creative prompts, generates a model image, and uses the original product and generated model as references for additional campaign visuals.

The final images are automatically uploaded to Google Drive, creating a streamlined process for producing consistent fashion and eCommerce assets without manually managing each stage of the creative workflow.

**Technologies:** n8n · Google Gemini · AI Image Generation APIs · Google Drive

## 01. The Problem: Manual Image Creation for Fashion Campaigns

Before automation, fashion campaign production required manual coordination across photography, creative direction, and AI image generation.

Common challenges included:

- Time-consuming image conceptualization and generation.
- Specialist prompt engineering for different image types.
- Fragmented workflows for managing product images.
- Delays and inconsistencies across campaign visuals.

**Core challenge:** Integrate the entire creative production pipeline into one automated system that generates high-quality, consistent fashion campaign images efficiently.

## 02. The Goal: A Fully Automated Fashion Photoshoot

The goal was to transform a single product photo into a complete set of AI-generated campaign assets through one workflow.

The system:

- Accepts a clothing image through a simple form.
- Analyzes the product and generates a suitable creative prompt.
- Creates a photorealistic model image featuring the product.
- Generates eCommerce studio shots and lifestyle campaign visuals.
- Uses the original clothing and generated model images as references.
- Uploads the resulting assets to Google Drive automatically.

This reduces manual coordination and streamlines creative production from input to final output.

## 03. The Solution: AI-Powered Fashion Image Generation

The workflow combines n8n automation with AI-powered image analysis, prompt generation, and image creation.

It uses internal workflow logic to develop creative directions, generates a model image from the original product photo, and then creates additional product and lifestyle visuals using both images as references.

All generated assets are organized in Google Drive, making them easier to access and manage.

## 04. How the System Works

### Phase 1: Image Input and Model Creation

1. A form trigger receives the uploaded clothing image.
2. The image is converted to Base64 for API processing.
3. Workflow logic analyzes the garment and develops a creative prompt, including the intended model persona.
4. An AI image generation API using Google Gemini creates the fashion model image.
5. The generated model image is uploaded to Google Drive.

### Phase 2: Prompt Generation and Campaign Image Creation

1. The original clothing image and generated model image are used as visual references.
2. The workflow creates prompts for eCommerce studio shots and lifestyle scenes.
3. AI image generation APIs produce the corresponding campaign visuals.
4. Generation results are checked for success.
5. Final images are uploaded to Google Drive.

## 05. Real-World Impact

- **Reduced Manual Prompting:** Creative direction and prompt generation are handled by the workflow.
- **Consistent Brand Visuals:** Products receive standardized eCommerce images and lifestyle scenes.
- **Faster Production:** Multiple stages that previously required manual coordination run automatically.
- **Scalable Content Creation:** The same workflow can be reused across a product catalog.
- **Organized Assets:** Generated images are stored in Google Drive with predictable naming conventions.

## 06. Expected Impact

### Time Savings

The workflow is designed to reduce manual creative and prompt-writing time by an estimated 70–90%, depending on the production process and level of review required.

### Consistency and Coverage

Standardized prompts help reduce missing image angles, incomplete product coverage, and inconsistent lifestyle visuals.

### Faster Content Production

Campaign visuals can potentially be generated on the same day products arrive, helping teams prepare new collections more quickly.

### Asset Organization

Generated images are uploaded to Google Drive using predictable naming conventions, such as `AI Model` and `Result #X`, making image sets easier to browse.

### Creative Efficiency

Reusable creative logic allows teams to produce more variations while reducing dependence on external photoshoots for every campaign asset.

## 07. Workflow Architecture

Image (https://mburhan.com/uploads/c86605bdf56b2ff7ed9b1e78.png)

- **Product Image Upload:** Receives the original clothing image.
- **Garment Analysis:** Extracts visual details to guide image generation.
- **Creative Prompt Generation:** Defines the model and visual direction.
- **AI Model Image:** Creates the initial model image using the product reference.
- **Campaign Image Generation:** Produces studio and lifestyle variations.
- **Google Drive:** Stores the generated assets for later use.

Image: Fig:n8n automation workflow (https://mburhan.com/uploads/43b28f623e6950ffc1b3440e.png)

## 08. Final Outcome

The workflow transforms a single clothing image into a collection of AI-generated fashion campaign visuals through an automated process.

- Creative prompting is handled within the workflow.
- Model, eCommerce, and lifestyle images are generated from visual references.
- Multiple stages of image production run automatically.
- Campaign assets are organized in Google Drive.
- Reusable automation supports production across larger product catalogs.

**The result:** A streamlined fashion content production pipeline that helps brands create campaign-ready visuals faster, with more consistent creative direction and less repetitive manual work.

---

# AI-Powered 24/7 Voice Support Agent for eCommerce Stores

> A 24/7 AI voice agent that answers store calls using your product and policy data, escalating only complex calls to humans.

Section: n8n Automations, AI Agents, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-24-7-voice-support

## Project Overview

The AI-Powered 24/7 Voice Support Agent automates customer support calls for eCommerce stores using conversational AI and a centralized knowledge base. Built around VAPI, n8n, Slack, and Google Sheets, it can answer common questions about products, pricing, shipping, returns, and store policies through natural voice conversations.

The system understands customer intent, retrieves relevant store information, responds in real time, and supports optional escalation for complex or high-risk issues. This reduces repetitive work for support teams while providing customers with access to assistance around the clock.

**Technologies:** VAPI · n8n · Slack · Google Sheets · AI Voice Agent

## 01. The Problem: Support Calls Don’t Scale

Before automation, customer support teams handled incoming calls manually.

Common challenges included:

- Repeating answers about products, sizes, shipping, returns, and order status.
- Switching between store systems and knowledge documents during calls.
- Missing calls during peak hours or outside business hours.
- Providing inconsistent answers depending on the agent handling the call.

**Impact per team**

- Long wait times and dropped calls.
- Frustrated customers seeking answers to basic questions.
- Inconsistent use of product and policy information.
- High dependency on human agents for repetitive FAQs.

**Core challenge:** Build a 24/7 support system that can resolve common customer queries using the store’s own knowledge while escalating issues that require human attention.

## 02. Results After Automation

After implementation, the voice agent can answer calls around the clock and handle common customer questions automatically.

**Key improvements**

- Calls answered without waiting for a human agent.
- Common FAQs handled through conversational AI.
- Complex or high-risk issues escalated when required.
- Responses grounded in available store product and policy information.
- Human agents able to focus on more complex customer needs.

**Time and effort savings**

**Before**

- Approximately 100–200 repetitive calls per week.
- Around 2–3 hours per day spent answering FAQs.

**After**

- Routine call handling automated.
- Human involvement primarily reserved for escalations and exceptions.

**Estimated monthly impact**

- Approximately 40–60 hours saved per support team.
- More time available for complex and high-value support cases.

**Consistency improvements**

- Responses use the configured product and policy knowledge base.
- Reduced reliance on improvised answers.
- Easier alignment with updated store information when the knowledge base is maintained.

## 03. The Solution: AI-Powered Voice Agent

The system acts as a virtual customer support representative capable of handling common eCommerce questions independently.

It:

- Provides natural, conversational voice responses 24/7.
- Uses store knowledge covering products, pricing, inventory, policies, and FAQs.
- Answers common questions without requiring a human agent.
- Maintains conversational context across follow-up questions.
- Supports optional escalation for complex or high-risk calls.
- Can send call summaries and capture relevant information through n8n, Slack, and Google Sheets.

## 04. How the System Works

### Phase 1: Knowledge Base and Store Data

Relevant store information is collected and organized for retrieval during calls.

This may include:

- Product details, sizes, colors, and availability.
- Pricing and promotions.
- Shipping fees and delivery times.
- Return, refund, and exchange policies.
- Frequently asked questions.

The knowledge base should be kept up to date to reduce the risk of outdated responses.

### Phase 2: Incoming Call and Intent Detection

1. A customer calls the support number.
2. VAPI answers with a configured greeting and voice.
3. The customer's speech is converted into text.
4. AI identifies the intent, such as a product question, order inquiry, or policy clarification.

### Phase 3: Knowledge Base Lookup

The system retrieves relevant information to answer the customer's question.

For example:

- Product specifications and availability.
- Delivery timelines and shipping costs.
- Return and refund conditions.
- Promotions and bundles.
- Order details for authenticated users, where the required integration is configured.

### Phase 4: Voice Response Generation

- AI prepares a relevant, conversational answer.
- Text-to-speech generates the spoken response.
- The answer is played to the customer in real time.

### Phase 5: Multi-Turn Conversation

The agent maintains context throughout the call, allowing customers to ask follow-up questions about products, orders, and policies without restarting the conversation.

### Phase 6: Human Escalation and Call Summary

When configured, complex or high-risk calls can be forwarded to a human agent. The workflow may also support callback requests or ticket creation.

After the call:

- VAPI sends call-end information to an n8n webhook.
- n8n processes the call summary and relevant details.
- Slack receives the summary.
- Google Sheets records applicable information.

## 05. Business Value

- **24/7 Availability:** Customers can receive assistance outside normal business hours.
- **Reduced Support Workload:** AI handles repetitive FAQs automatically.
- **Consistent Answers:** Responses are based on configured store knowledge.
- **Faster Resolution:** Common questions can be answered in a short conversation.
- **Better Knowledge Utilization:** Product and policy information supports customer interactions.
- **Human Escalation:** Complex cases can be routed to people when required.
- **Call Visibility:** Summaries and captured details can be recorded for follow-up.

## 06. Expected Impact

### Time Savings

Automating repetitive calls reduces the time support agents spend answering routine questions, allowing them to focus on complex issues.

### Error Reduction

Using maintained product and policy data helps reduce inconsistent or improvised answers.

### Call Handling Speed

Common questions can potentially be resolved in one or two exchanges, depending on their complexity.

### Knowledge Utilization

Centralized store information can be reused across voice interactions, improving consistency and accessibility.

**Estimated time savings:** Approximately 40–60 hours per month per support team, depending on call volume and automation coverage.

## 07. Workflow Architecture

Image (https://mburhan.com/uploads/441b8a39a9086aab67b8de19.png)

The main components work together as follows:

- **VAPI:** Handles voice interactions and call management.
- **Knowledge Base:** Supplies relevant product and policy information.
- **AI Processing:** Understands customer questions and prepares responses.
- **n8n:** Processes call-end events and coordinates follow-up actions.
- **Slack:** Delivers call summaries to the support team.
- **Google Sheets:** Stores applicable call records and captured information.

## 08. How to Replicate the System

### Step 1: Configure the VAPI Voice Agent

- Create the voice agent and configure its greeting and voice tone.
- Connect the store knowledge base.
- Test its ability to answer FAQs and product or policy questions.

### Step 2: Configure the Call-End Webhook

- Create a webhook node in n8n.
- Configure VAPI to send call-end information to the webhook.
- Verify that the webhook receives the expected call details.

### Step 3: Import the n8n Workflow

1. Open `ShopTalk AI – Call Summary & Lead Capture (1).json`.
2. Import the workflow into n8n.
3. Connect the VAPI webhook to the workflow.
4. Configure Slack and select the appropriate summary channel.
5. Connect Google Sheets and select the spreadsheet and required columns.

### Step 4: Test the Flow

- Make a test call.
- Ask product or policy questions.
- Test a follow-up question.
- End the call.
- Verify that the call summary appears in Slack.
- Confirm that a row is added to Google Sheets when applicable.

### Step 5: Activate the Workflow

Once testing is complete, activate the n8n workflow and verify that the VAPI webhook is configured correctly.

The voice agent can then handle customer calls continuously, while n8n processes call summaries and related follow-up actions.

## 09. Final Outcome

The AI-powered voice support system transforms repetitive customer support calls into a more consistent and scalable operation.

- Customers can receive answers at any time.
- Human agents spend less time handling repetitive FAQs.
- Product and policy information supports consistent responses.
- Complex issues can be escalated when necessary.
- Call summaries and relevant records improve follow-up visibility.
- Support teams can handle routine demand without relying exclusively on additional staff.

**The result:** A scalable, 24/7 voice support experience that reduces repetitive work while keeping human agents available for the issues that need them most.

---

# AI-Powered Social Media Content Creation

> An automation that generates LinkedIn captions and images, sends them for approval, and publishes only approved posts.

Section: n8n Automations, AI Agents, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-social-media-content-creation

## Project Overview

The AI-Powered Social Media Content Creation system automates the process of generating, reviewing, and publishing LinkedIn content. Built with n8n, LinkedIn, Google Sheets, and Google Drive, it uses AI to create captions and matching post images while maintaining a centralized record of content and approval status.

By introducing an approval step before publishing, the system helps teams maintain consistent brand messaging, reduce repetitive work, and manage their social media content through a single workflow.

**Technologies:** n8n · LinkedIn · Google Sheets · Google Drive · AI Content Generation

## 01. The Problem: Social Media Management Is Inconsistent

Before automation, creating and publishing social media content required multiple tools and manual steps.

Teams had to:

- Write captions repeatedly.
- Design post images manually.
- Maintain content variety and avoid duplication.
- Track approvals through email or chat.
- Publish posts manually on LinkedIn.
- Keep track of drafts, approvals, and published content.

**Impact per team**

- 2–4 hours per day spent creating, approving, and publishing posts.
- Risk of duplicate, missed, or unapproved posts.
- Inconsistent brand messaging across channels.

**Core challenge:** Teams needed a reliable system to create, review, approve, and publish content at scale without unnecessary operational complexity.

## 02. Results After Automation

After implementation, content creation and publishing became a structured, automated process.

**Key improvements**

- Captions and images are generated automatically.
- Posts are published only after approval.
- Google Sheets provides a centralized content control panel.
- Manual content creation and publishing effort is significantly reduced.

**Time and effort savings**

**Before**

- Approximately 3 hours per day.
- Around 60 hours per month, based on 20 working days.

**After**

- Automated content generation and publishing.
- Human involvement mainly required for approvals, estimated at 5–10 minutes.

**Monthly impact**

- Approximately 50–55 hours saved per team per month.
- More time available for planning, strategy, and other creative work.

**Consistency and tracking improvements**

- Approved posts follow the configured brand tone and guidelines.
- Approval controls help prevent unauthorized publishing.
- Content status and live post URLs are recorded centrally.

## 03. The Solution: AI-Powered Social Media Automation

The system acts as a virtual social media assistant that manages content from creation to publication.

It:

- Generates LinkedIn captions based on topics, brand tone, and content goals.
- Creates post images aligned with the generated captions.
- Uses Google Sheets to manage content and approval status.
- Sends posts for approval before publishing.
- Publishes approved posts to LinkedIn.
- Records publishing status and live URLs for tracking and reporting.

## 04. How the System Works

### Phase 1: AI Caption Generation

The workflow starts when a new row is added to Google Sheets.

AI generates captions based on:

- Topic.
- Brand context.
- Content objective.

The captions are formatted for LinkedIn and saved in Google Sheets for review and approval.

### Phase 2: Image and Post Creation

The generated caption is used to create a matching AI-generated image.

- The image is uploaded to Google Drive.
- The image URL is stored in Google Sheets.
- Caption, image URL, and status are maintained in the control panel.
- Content remains visible for review and approval.

### Phase 3: Approval and Publishing

- Approval emails are sent automatically.
- Posts remain blocked until approval is received.
- Only approved posts are published on LinkedIn.

Google Sheets is updated with:

- Publishing status.
- Live post URL.

This provides a centralized record of the content lifecycle and helps prevent missed or duplicate publishing.

## 05. Business Value

- **End-to-End Automation:** Manages content generation through publishing in one workflow.
- **Centralized Control Panel:** Tracks drafts, approvals, and published posts.
- **Consistent Brand Voice:** Applies configured tone and style guidelines.
- **Approval-Safe Publishing:** Prevents publishing before explicit approval.
- **Full Tracking and Reporting:** Maintains content statuses and live URLs.
- **Time and Cost Savings:** Saves an estimated 50+ hours per team each month.

## 06. Workflow Architecture

Image (https://mburhan.com/uploads/78d804c57345629f032a041a.png)

Each component has a defined role:

- **Google Sheets:** Triggers content creation and tracks captions, image URLs, and statuses.
- **AI Caption Generation:** Creates captions from the topic and brand context.
- **Image Generation:** Produces visuals aligned with the captions.
- **Google Drive:** Stores generated images.
- **Approval:** Ensures posts are reviewed before publishing.
- **LinkedIn Publishing:** Publishes approved content.
- **Status Update:** Records publication status and the live post URL.

## 07. How to Replicate the System

### Step 1: Import the Workflow

1. Open the attached `Social Media Automation Workflow.json` file.
2. Copy the complete JSON content.
3. Open your n8n cloud or self-hosted instance.
4. Create a new workflow and paste or import the JSON.
5. Save and rename it to **Social Media Content Automation**.

### Step 2: Configure Connections

- **Google Sheets:** Connect your account and configure columns for Topic, Caption, Image URL, Status, and Post URL.
- **AI Caption Generation:** Add the required AI credentials and verify the brand context and prompts.
- **Image Generation:** Confirm that image prompts are derived from captions and images are saved to Google Drive.
- **Approval:** Configure email credentials and verify the approval logic.
- **LinkedIn:** Connect the appropriate account and configure publishing permissions.

### Step 3: Test the Workflow

Run the workflow manually and verify that:

- Captions and images appear in Google Sheets.
- Approval logic blocks unapproved posts.
- Approved posts publish correctly.
- Publishing statuses and live URLs update accurately.

### Step 4: Activate the Automation

Once testing is complete, activate the workflow and verify its trigger settings.

Content generation, approval, and publishing can then run automatically with minimal manual involvement.

## 08. Final Outcome

The AI-powered workflow transforms social media management into a consistent and trackable process.

- Content creation becomes faster and more repeatable.
- Publishing follows an approval-based process.
- Teams manage content through a centralized control panel.
- Publishing activity and live URLs remain available for tracking.
- An estimated 50+ hours of manual work can be saved per team each month.

**The result:** A more efficient, approval-controlled social media workflow that allows teams to scale content production without proportionally increasing manual effort.

---

# AI-Powered E-Commerce Product Price Monitoring

> A scheduled workflow that tracks product prices, logs history to Google Sheets, and alerts you only on meaningful price drops.

Section: n8n Automations, AI Agents, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-e-commerce-product-price

## Project Overview

The AI-Powered E-Commerce Product Price Monitoring system automates product price tracking, historical data collection, and intelligent price-drop alerts. Built with n8n, Google Sheets, Slack, OpenAI, and Gemini, it regularly collects product prices, compares them with historical records, and identifies meaningful changes.

By automating repetitive price checks and notifying users only when relevant opportunities arise, the system helps users make informed purchasing decisions while saving time and avoiding unnecessary alerts.

**Technologies:** n8n · Google Sheets · Slack · OpenAI · Gemini

## 01. The Problem: Manual Price Checking Wastes Time

Before automation, tracking product prices across online stores was slow, repetitive, and error-prone.

Users had to manually:

- Open multiple product pages.
- Check prices repeatedly.
- Record or remember previous prices.

This often resulted in missed price drops, limited-time deals, and a lack of historical visibility into price trends. Buying decisions were frequently based on incomplete information rather than reliable data.

**Impact per user**

- 30–60 minutes per day spent manually checking prices.
- Missed savings on deals and discounts.
- Decisions based on incomplete price information.

**Core challenge:** Users want to buy at the right time without constantly monitoring product prices themselves.

## 02. Results After Automation

After implementation, the system automatically monitors prices, records historical data, and sends alerts when meaningful changes occur.

**Key improvements**

- Automated daily price monitoring.
- Notifications triggered by meaningful price changes.
- Historical price trends available in Google Sheets.
- Data-driven purchasing decisions.
- Minimal human involvement in routine monitoring.

**Time and effort savings**

**Before**

- Approximately 45 minutes per day.
- Around 22.5 hours per month spent checking prices manually.

**After**

- Price checks and comparisons run automatically.
- Human effort for routine monitoring is close to zero.
- Approximately 20–25 hours saved per user each month.

**Accuracy and opportunity gains**

- Consistent recording of monitored prices.
- Timely alerts for price drops and limited-time deals.
- Better comparisons based on historical data and AI analysis.

## 03. The Solution: AI-Powered Price Monitoring

The system acts as a virtual price tracker and analyst by combining scheduled data collection, historical tracking, AI analysis, and automated notifications.

It:

- Collects product prices on a defined schedule.
- Logs historical prices in Google Sheets.
- Evaluates price changes and potential value opportunities using AI.
- Sends notifications through Slack or other configured channels.
- Reduces notification noise by alerting users only when meaningful changes occur.

## 04. How the System Works

### Phase 1: Scheduled Price Collection

The workflow runs automatically on a daily or custom schedule.

It visits configured product pages and extracts:

- Product name.
- Current price.

Incomplete or erroneous entries are skipped.

### Phase 2: Price Comparison and History Logging

- Current prices are compared across monitoring runs.
- Each successful run adds a new row to Google Sheets.
- Historical records build a timeline of price changes.

### Phase 3: AI-Driven Analysis

OpenAI or Gemini evaluates the collected information to help identify:

- Current prices compared with historical prices.
- Products offering better value.
- Significant changes that may justify a notification.

### Phase 4: Smart Notifications

Notifications are sent when configured conditions are met, such as a significant price drop or a better deal becoming available.

- Alerts can be delivered through Slack, email, or other configured channels.
- No alert is sent when there is no meaningful change.

## 05. Business Value

- **Zero Manual Monitoring:** Routine price checks run automatically.
- **Never Miss a Deal:** Timely alerts highlight potential savings.
- **Data-Driven Decisions:** AI analysis supports product comparisons.
- **Historical Visibility:** Price records reveal trends over time.
- **Reduced Notification Noise:** Users receive fewer unnecessary alerts.
- **Time Savings:** Approximately 20–25 hours saved per user each month.

## 06. Workflow Architecture

The automation follows this sequence:

Image (https://mburhan.com/uploads/82dcb7ad69950c413abf79b1.png)

Each component has a specific role:

- **Scheduler:** Starts the workflow at configured intervals.
- **Price Collection:** Retrieves product details and current prices.
- **Price Comparison:** Identifies differences between current and previous records.
- **Google Sheets:** Stores historical price data.
- **AI Analysis:** Evaluates price trends and potential opportunities.
- **Smart Notification:** Sends alerts when configured conditions are met.

## 07. How to Replicate the System

### Step 1: Import the Workflow

1. Open the attached `E-Commerce Price Monitoring Automation.json` file.
2. Copy its complete JSON content.
3. Open your n8n cloud or self-hosted instance.
4. Create a new workflow and import or paste the JSON.
5. Rename the workflow to **E-Commerce Price Monitoring Automation**.

### Step 2: Configure Connections

- **Scheduler:** Set the monitoring frequency.
- **HTTP or scraping nodes:** Configure product URLs and test price extraction.
- **Google Sheets:** Connect your account and map the Product, Price, Date, and Comparison Result columns.
- **AI analysis:** Configure OpenAI or Gemini credentials.
- **Notifications:** Connect Slack or email and configure alert messages.

### Step 3: Test the Workflow

Run the workflow manually and verify that:

- Product data is collected correctly.
- Price history is recorded in Google Sheets.
- AI analysis evaluates the collected data.
- Notifications are sent when the configured conditions are met.
- No alert is sent when prices remain unchanged and no other alert condition is triggered.

### Step 4: Activate the Automation

Once testing is complete, activate the workflow and confirm the schedule is configured correctly.

The system will then monitor product prices, maintain historical records, analyze changes, and send alerts according to the configured rules.

Image (https://mburhan.com/uploads/9a25b5d40f6ecf6f18d7c089.png)

## 08. Final Outcome

The AI-powered price monitoring workflow transforms repetitive, manual price checking into an automated and data-driven process.

Users can monitor multiple products, review historical prices, identify potential savings, and receive meaningful alerts without continuously checking online stores.

**The result:** Less manual work, better price visibility, and more informed purchasing decisions.

---

# AI-Powered Shopify Abandoned Cart Recovery System

> An always-on automation that detects abandoned Shopify carts and sends personalized, AI-written recovery emails.

Section: n8n Automations, Make Automations, AI Agents, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-shopify-abandoned-cart-recovery

## 01. The Problem: Abandoned Carts Were Leaving Revenue on the Table

Before automation, abandoned carts were often handled manually or received no follow-up at all.

- Teams had to check Shopify manually for abandoned checkouts.
- Writing reminder emails took time and effort.
- Follow-ups were inconsistent, and many carts were never contacted.
- Generic emails lacked personalization.
- Delayed reminders reduced the chance of customers completing their purchases.

### The Cost of Manual Follow-Ups

For stores handling 40–60 abandoned carts per day:

- Each follow-up took approximately 5–7 minutes.
- Manual work could consume 3–5 hours daily.
- Some stores followed up only once a day or skipped follow-ups entirely.

Without timely reminders, potential sales were lost.

## 02. Results After Automation

The automated workflow ensures abandoned checkouts receive timely, personalized recovery emails with minimal human intervention.

- **100% recovery-attempt coverage:** Every detected abandoned cart receives a recovery attempt.
- **90–95% less manual work:** Repetitive follow-up tasks are automated.
- **Faster communication:** Recovery emails are sent within the configured time window.
- **AI personalization:** Each email is generated to provide a relevant, friendly reminder.

### Time Savings

| Metric | Before Automation | After Automation |
| --- | --- | --- |
| Follow-up process | Manual | Automated |
| Time per cart | Approximately 6 minutes | Minimal manual effort |
| Daily workload | Around 3–5 hours | Near-zero routine effort |
| Monthly time savings | — | Approximately 100–120 hours |

### Potential Revenue Impact

Using a conservative example of 50 abandoned carts per day, a $50 average cart value, and a 5–8% recovery rate:

- **Potential recovered sales:** $125–$200 per day.
- **Potential recovered sales:** Approximately $3,750–$6,000 per 30-day month.

These figures are illustrative estimates, not guaranteed results. Actual performance depends on store traffic, customer behavior, email deliverability, and conversion rates.

## 03. The Solution: AI-Powered Abandoned Cart Recovery

The system works as an always-on sales assistant that follows up with customers automatically.

It:

- Detects abandoned Shopify checkouts.
- Waits for a configured period before sending a reminder.
- Verifies that the checkout is still abandoned and has not been completed.
- Uses AI to generate a personalized recovery email.
- Sends the email automatically through Gmail or a configured email service.

The workflow operates in the background, reducing repetitive work while keeping follow-ups consistent.

## 04. How the Automation Works

The workflow connects Shopify, an automation platform, AI, and email delivery.

1. **Abandoned Checkout Detection:** Shopify provides abandoned checkout data.
2. **Wait and Verify:** The automation waits and checks whether the checkout remains abandoned.
3. **AI Email Generation:** OpenAI generates a friendly, personalized recovery message.
4. **Email Delivery:** Gmail or the configured email service sends the message to the customer.
5. **Scheduled Execution:** The workflow runs automatically according to its configured schedule.

## Workflow:

Image (https://mburhan.com/uploads/b7ebb79125925b2fb319880b.png)

No routine manual follow-up is required once the workflow is configured and activated.

## 05. Business Value

### Recover Potential Revenue

Re-engages customers who leave products in their carts and creates another opportunity to complete a purchase.

### Reduce Manual Work

Eliminates repetitive checkout monitoring and email-writing tasks.

### Improve Customer Experience

Delivers timely reminders with more relevant, personalized messaging.

### Ensure Consistent Follow-Ups

Provides a recovery attempt for every detected abandoned checkout that meets the workflow conditions.

### Scale With Store Growth

Handles increasing checkout volumes without requiring a proportional increase in manual effort.

## 06. How to Replicate the System

Use the provided workflow blueprint to configure the automation in Make.com.

### Step 1: Import the Blueprint

1. Open the attached JSON file: **Shopify Abandoned Checkout Recovery with Order Verification (1).json**.
2. Select and copy the complete JSON content.
3. Log in to Make.com and open **Scenarios**.
4. Create a new scenario and choose the blueprint import option.
5. Import the JSON, save the scenario, and name it **Shopify Abandoned Checkout Recovery with Order Verification**.

### Step 2: Configure Shopify

- Connect your Shopify account.
- Confirm that the workflow retrieves abandoned checkouts.
- Verify that the checkout status and required customer fields are mapped correctly.

### Step 3: Configure the Wait and Verification Logic

- Review the delay before sending the recovery email.
- Verify that the checkout has not been completed.
- Prevent duplicate or unnecessary recovery emails.

### Step 4: Configure OpenAI

- Add your OpenAI API key.
- Confirm that the AI generates a relevant, friendly recovery email.
- Check that the email body follows your desired format and tone.

### Step 5: Configure Email Delivery

- Reconnect Gmail or your SMTP email service.
- Map the customer's email address dynamically.
- Set the email subject to **You left something behind**.
- Verify the recipient, subject, and generated email body.

### Step 6: Test and Activate

1. Run the scenario once.
2. Verify that abandoned checkout data is retrieved.
3. Confirm that the checkout verification works.
4. Check that the AI-generated email is correct.
5. Confirm that the recovery email is delivered successfully.
6. Activate the scenario and configure an appropriate schedule, such as every 15–60 minutes.

**Important:** The actual schedule, delay, and verification rules should match your store's requirements and the imported workflow configuration.

## 07. Final Outcome

The AI-Powered Shopify Abandoned Cart Recovery System transforms abandoned checkouts into automated follow-up opportunities.

By combining Shopify checkout data, workflow automation, AI-generated emails, and automated delivery, the system helps businesses:

- Follow up consistently with abandoned carts.
- Reduce repetitive manual work.
- Deliver timely, personalized reminders.
- Create additional opportunities to recover lost sales.
- Scale their recovery process as the business grows.

**The result:** A reliable, always-on abandoned cart recovery workflow that supports revenue recovery without increasing routine workload.

---

# AI-Powered Whatsapp message Assistant

> An AI-powered WhatsApp assistant that automates customer support through text, voice, and image understanding, delivering instant, knowledge-based responses 24/7.

Section: n8n Automations, AI Agents, AI Agents for WhatsApp, AI Agents for Ecommerce

Page: https://mburhan.com/case/ecom-agent-ai-powered-whatsapp-message-assistant

## 01.Project Overview

The AI-Powered WhatsApp Message Assistant is an intelligent customer support solution built with n8n, OpenAI, WhatsApp, and Supabase. It automates customer conversations by understanding text, voice messages, and images, then generating relevant responses based on business knowledge.

Using AI-powered processing and semantic search, the assistant delivers consistent, context-aware support around the clock, reducing repetitive manual work and allowing human agents to focus on more complex customer inquiries.

## 02. The Problem: WhatsApp Customer Support Didn’t Scale

Before automation, customer support over WhatsApp was entirely manual and increasingly difficult to manage.

Staff had to respond to every message individually, including repetitive questions about products, prices, availability, and store policies. Voice messages required manual listening and interpretation, while image-based questions, such as product photos, needed human review. Responses varied depending on the staff member, and support was unavailable outside business hours.

For a small business receiving **80–120 WhatsApp messages per day**, this resulted in:

- **High Manual Workload:** 4–6 hours per day spent answering routine questions.
- **Delayed Responses:** Longer waiting times during peak hours.
- **Missed Inquiries:** Customer messages going unanswered outside working hours.
- **Staff Dependency:** Heavy reliance on trained employees with product knowledge.
- **Inconsistent Support:** Different responses depending on who handled the conversation.

**Core Challenge:** WhatsApp was the main customer communication channel, but human-only support could not keep up with demand without increasing headcount.

---

## 03. Results After Automation

After deploying the WhatsApp AI Business Assistant:

- **70–85% of incoming queries** were handled automatically.
- Customers received instant replies, including outside business hours.
- Voice messages were transcribed and answered automatically.
- AI understood image-based queries involving products and issues.
- Responses became more consistent, accurate, and aligned with business policies.
- Human agents focused on complex or exceptional conversations.

### Time and Workload Impact

**Before Automation**

- Approximately 3–4 minutes per WhatsApp message.
- Around 100 messages per day required approximately 5–6 hours of support time.

**After Automation**

- Human involvement was limited to approximately 15–30 messages per day.
- Manual support time decreased to around 1–1.5 hours per day.

**Estimated Time Savings**

- 3–4 hours saved per day.
- 90–120 hours saved per month.

This allowed the business to scale WhatsApp support without hiring additional staff.

---

## 04. The Solution: An AI Assistant Inside WhatsApp

The system acts as a virtual support agent directly within WhatsApp.

- Understands text, voice messages, and images.
- Maintains conversation context.
- Retrieves relevant business knowledge instantly.
- Generates natural, context-aware responses.
- Supports text or voice replies when appropriate.

Customers receive fast, helpful support without needing to change how they communicate with the business.

---

## 05. How the System Works

1. **Customer Message:** Customers contact the business through WhatsApp.
2. **Message Understanding:** AI identifies the message type and intent.
3. **Knowledge Retrieval:** Relevant business information is retrieved using semantic search.
4. **Response Generation:** AI generates a context-aware answer based on the available information.
5. **WhatsApp Reply:** The response is delivered automatically to the customer.

The workflow processes messages automatically within seconds.

Image (https://mburhan.com/uploads/2b82f4098943c7849229055c.png)

---

## 06. Business Value

- **Reduced Manual Workload:** Automates repetitive customer support tasks.
- **Faster Responses:** Helps customers receive timely answers.
- **Consistent Communication:** Maintains more uniform, policy-aligned responses.
- **Multilingual Support:** Handles conversations in different languages.
- **24/7 Availability:** Supports customers beyond regular business hours.
- **Seamless Experience:** Customers continue using WhatsApp as usual.

---

## 07. Reliability and Safety

- **Business-Grounded Answers:** Uses available business knowledge to guide responses.
- **Graceful Fallbacks:** Handles unclear information without relying on unsupported answers.
- **Data Boundaries:** Does not process payments or sensitive data.
- **Human Escalation:** Complex cases can be directed to human agents when escalation is configured.

These measures help maintain reliable, consistent, and responsible customer support.

---

## 08. How to Replicate the System

The solution uses the **WhatsApp Cloud API, n8n, OpenAI, and Supabase Vector Store**.

### Step 1: Set Up WhatsApp Cloud API

Configure the WhatsApp Business account, phone number ID, and webhook events for incoming messages.

Ensure the integration supports receiving text, voice, and image messages and sending text or audio replies.

### Step 2: Create the WhatsApp Webhook in n8n

- Create a Webhook Trigger node in n8n.
- Copy the generated webhook URL.
- Add the URL to the Meta WhatsApp Webhooks configuration.

This connects incoming WhatsApp messages to the automation workflow.

### Step 3: Import the n8n Workflow

Create a new workflow in n8n and import the provided file:

`AI-Powered WhatsApp Assistant.json`

### Step 4: Connect AI and Data Services

**OpenAI**

- Add the OpenAI API key.
- Configure the chat model, audio transcription, image analysis, and text-to-speech capabilities.

**Supabase Vector Store**

- Add the Supabase project URL and API key.
- Verify that the vector table contains products, FAQs, policies, and business information.

This enables retrieval-augmented generation (RAG) for business-grounded responses.

### Step 5: Configure Message Routing

Open the Switch node and verify the routing rules:

- **Text:** Text message → AI Agent
- **Voice:** Voice message → Transcription → AI Agent
- **Image:** Image message → Image Analysis → AI Agent

This separates message-processing paths according to the input type.

### Step 6: Test the Workflow

Send test messages through WhatsApp, including:

- A text question, such as a product inquiry.
- A voice message.
- An image, such as a product photo.

Verify correct intent detection, relevant answers from business knowledge, successful WhatsApp replies, and voice responses when triggered.

### Step 7: Activate the Workflow

Switch the n8n workflow to **Active** after testing and validating the integrations.

The assistant can then respond to customer messages automatically through WhatsApp.

---

## 09. Final Outcome

With the WhatsApp AI Business Assistant deployed:

- Customer support became faster and more reliable.
- Staff could focus on higher-value conversations.
- Customers received timely, relevant assistance.
- Support could scale with business growth without proportionally increasing manual workload.

The system transformed WhatsApp from a manual support burden into an always-on, AI-powered customer assistant.
