IBADAT ALI
AI Automation Engineer | n8n & AI Agents,bussiness workflows
- 1
- Followers
Recently I completed an AI automation project for a client who was running payment operations through WhatsApp 💬
Here's how it worked before I got involved: customers would pay via Stripe, screenshot their receipt, and send it to a WhatsApp group. Someone on the team would open Stripe, manually search for that transaction, and type a reply in the chat. They had 50+ active groups. This was happening hundreds of times a day 😅
No system to catch fake receipts. No system to catch someone sending the same screenshot twice. Just humans, doing the same lookup over and over.
So I built a bot that sits inside every group and handles it automatically 🤖
When a payment screenshot comes in, it extracts the transaction data using OCR (Tesseract locally, Groq Vision or Gemini as fallback for low-quality images), checks it against a duplicate store so recycled receipts don't slip through, then hits the Stripe API to verify the payment is real. The primary match key is the customer's email from the WhatsApp caption not just the transaction ID, because that alone is too easy to fake. It replies
directly in the group with the verdict: ✅ verified, 🚨 fraud flagged, or 🔁 duplicate.
Everything logs to Google Sheets. Anything suspicious fires a Telegram alert.
The whole thing runs as three Docker services on a DigitalOcean VPS, 24/7 ⚙️ WhatsApp bot in Node.js, OCR microservice in Python/FastAPI, and n8n handling the verification workflow.
If a service crashes, it restarts automatically.
One thing I was deliberate about: the system always fails toward "unverified," never toward
"verified." Low-confidence OCR goes to human review, not to an auto-approval. That decision
matters a lot when you're touching payments 🔐
The client no longer has anyone manually opening Stripe. Fakes and duplicates get caught
before anyone acts on them 🎯:
0
25
AI-Powered WhatsApp Restaurant Assistant — n8n + Google Gemini ,Google Sheets (FAQ, Inventory, Orders Automation)!
🍽️ AI-Powered WhatsApp Restaurant Assistant
n8n + Google Gemini + Google Sheets
(Project Case Study)
Built a structured AI assistant that runs inside WhatsApp and
handles real restaurant operations — answering questions,
checking inventory, and recording orders automatically.
One message in. A complete restaurant workflow behind it.
⚡ WHAT THE ASSISTANT HANDLES:
❓ FAQ Requests
Customer asks a question via WhatsApp
AI Agent reads the FAQ data source
Instant accurate reply sent back
📦 Inventory Availability Checks
Customer asks if an item is available
AI Agent queries real-time inventory tool
Replies with live stock status
📋 Order Placement
Customer sends their order via WhatsApp
AI Agent structures and validates the request
Order Tool records it in the system
Confirmation reply sent instantly
🧠 HOW THE SYSTEM IS DESIGNED:
🤖 AI Agent acts as the central reasoning layer (Understand, Decide, Use Tools)
💬 Simple Memory maintains conversation context across the session
🛠️ Three structured tools handle all restaurant actions: FAQ Tool, Inventory Tool, Order Tool
✅ AI generates the reply only after tools return results — not from assumptions
🔀 Multi-intent queries handled in a single agent loop
📊 WHAT THIS SYSTEM DELIVERS:
⚡ Faster Replies: Instant responses on WhatsApp
🧠 Smarter Handling: Understands questions, items, and orders using AI reasoning
📋 Better Operations: FAQs, inventory, and orders organized in one system
👥 Scalable Support: Handles multiple customers simultaneously without extra staff
🛠️ TECH STACK:
⚙️ Workflow Orchestration: n8n
💬 Messaging Platform: WhatsApp (Webhook)
✨ AI Reasoning: Google Gemini Chat Model
🧠 Memory: Simple Memory (conversation context)
📊 Restaurant Data Backend: Google Sheets (FAQ, Inventory, Orders)
🤖 Agent Architecture: AI Agent with structured tool calls
✅ DELIVERABLES:
✅ Fully automated WhatsApp assistant for restaurant ops
🍽️ Three operational tools: FAQ, Inventory, Orders
🧠 Context-aware conversation with session memory
📋 Real-time order recording in Google Sheets
👥 Handles concurrent customer conversations
🤖 AI that works through structured tools, not freeform responses
🔗 LINKS:
🌐 Portfolio: https://ibadat-ali-portfolio.vercel.app/
"My work is done when you no longer need me." — Ibadat Ali
0
58
🤖 AI-Powered Lead Generation System
n8n + Google Maps + Gemini AI Agent + Gmail Automation
(End-to-End, Private Client)
Built a complete AI-powered lead generation system for a private
client — from a single form submission to qualified prospects,
personalized outreach, and automated reporting. No manual steps.
⚡ THE PIPELINE (10 STEPS):
1️⃣ 📋 Web form submission triggers the entire workflow
2️⃣ 🗺️ Google Maps API discovers relevant businesses automatically
3️⃣ 🌐 Website scraper extracts live company information
4️⃣ 🚫 Duplicate detection removes repeated or known prospects
5️⃣ ⚙️ Business logic qualifies and scores each lead before
any AI call is made
6️⃣ 💾 Qualified leads stored in Google Sheets with full
state tracking
7️⃣ 🔍 Website research fetches real company context
for enrichment
8️⃣ ✨ Google Gemini AI Agent generates context-aware,
personalized outreach emails
9️⃣ ✉️ Gmail sends each email automatically after
content validation
🔟 📊 Completion report is built and emailed to the
client after every run
💡 DESIGN DECISIONS WORTH NOTING:
🧠 AI runs after business logic, not before it (fewer API calls, lower cost, more predictable output)
📋 Every lead state is recorded in Sheets — nothing is silently dropped
🔀 Fallback branches handle missing websites and emails
⏱️ Random delay added for rate-safe scraping
📩 Client receives a summary report, not raw workflow logs
🛠️ TECH STACK:
⚙️ Workflow Orchestration: n8n
🗺️ Lead Discovery: Google Maps API
🌐 Enrichment: Website Scraping and Content Extraction
✨ AI Reasoning: Google Gemini AI Agent
✉️ Outreach: Gmail (automated and validated before send)
📊 Storage and Reporting: Google Sheets
✅ DELIVERABLES:
✅ End-to-end lead operations pipeline (not a demo)
🎯 AI-personalized outreach — context-aware, not template-driven
🔍 Duplicate detection and lead qualification built in
📊 Automated completion report after every run
📋 Full lead state tracking across every pipeline stage
🔗 LINKS:
🌐 Portfolio: https://ibadat-ali-portfolio.vercel.app/
"My work is done when you no longer need me." — Ibadat Ali
0
72
WhatsApp Payment Verification AI Agent — n8n + OCR + Stripe + FastAPI (Production, Live Client)
Built and deployed a production-grade WhatsApp AI Agent that automatically
verifies payment screenshots against live Stripe data — no manual review,
no false approvals, running 24/7.
This was delivered for a real client and is currently live on DigitalOcean.
HOW IT WORKS:
Customer sends a payment screenshot via WhatsApp
Agent downloads and runs OCR (Tesseract) to extract: Amount, Ref ID, Date, Status
Extracted data is verified against live Stripe API — not the screenshot alone
Agent returns one of three decisions instantly via WhatsApp: ✅ Approved — Stripe confirmed exact match ⚠️ Requires Review — no exact match found ❌ Not Approved — verification rules not satisfied
FRAUD RESISTANCE BUILT IN:
Duplicate detection via perceptual image hashing (DUPLICATE_IMAGE_PHASH_RECEIPT_EVIDENCE) — the same screenshot cannot be reused for a second approval
Safe-fail design: visually convincing receipts are NOT auto-approved unless Stripe confirms them
Idempotency at the message level prevents replay attacks
PRODUCTION STACK:
WhatsApp: Node.js + Baileys
Backend & OCR: Python + FastAPI + Tesseract
Workflow Automation: n8n
Payment Verification: Stripe API (Live mode)
Deployment: Docker + DigitalOcean (24/7)
DELIVERABLES:
✔ Fully automated, end-to-end payment verification pipeline
✔ Three-state decision engine (Approved / Review / Not Approved)
✔ Duplicate screenshot detection
✔ Deployed and running in production — not a demo
Portfolio: https://ibadat-ali-portfolio.vercel.app
(https://ibadat-ali-portfolio.vercel.app)
"My work is done when you no longer need me." — Ibadat Ali
0
65