Automating KYC Verification for a Fintech Company with AI, APIs and Make
Manual identity verification is one of the biggest bottlenecks in fintech. It's time-consuming, inconsistent, and slows down user onboarding — often resulting in churn and compliance risks.
In this case study, we show how we helped a fast-growing fintech automate its KYC (Know Your Customer) process using Make.com, OCR + AI tools, and multi-layered API validation, cutting manual review by 93% while improving accuracy and speed.
The Problem
The client — a B2C fintech startup offering digital wallets and payment cards — struggled with scaling their KYC process:
- Each user upload (passport, ID, selfie) was reviewed manually
- Compliance officers manually validated identity, address, and face match
- High turnaround time (up to 48 hours)
- Many users dropped out before verification completed
- Limited audit trail for compliance reviews
The Goals
- Cut verification time to under 5 minutes
- Automate 80–90% of all standard KYC cases
- Ensure GDPR compliance and full audit trail
- Improve conversion rates for new sign-ups
- Minimize internal workload without compromising quality
Our Automation Solution
We built a robust verification flow using:
- Make.com — for orchestration of logic and multi-step scenarios
- OCR & facial recognition APIs — to validate ID and match selfie
- Government-issued ID verification — via third-party KYC API
- Airtable — for tracking progress, statuses, and exceptions
- OpenAI — to auto-flag anomalies in metadata or formatting
- Slack — for alerts to compliance staff in edge cases
How It Works
1. Trigger: User Uploads Verification Docs
- Form or app submission sends a webhook to Make
- Uploaded images (ID + selfie) passed to OCR + facial match API
2. AI-Powered Identity Checks
- OCR parses ID data: name, DOB, country, doc number
- Face comparison API matches selfie vs ID photo
- Metadata (e.g. blur, glare, tampering) scored with AI rules
3. Backend Validation & Country Match
- Compare declared country vs document issuer
- Validate document ID against third-party database
- Geo-IP check of session for fraud flagging
4. Result Routing & Slack Escalation
- ✅ If clean: auto-approved, user notified, stored in Airtable
- ❌ If flagged: routed to Slack channel for manual review
- All actions logged in Airtable with timestamps
Results After Implementation
| Metric | Before Automation | After Automation |
|---|---|---|
| Avg. Verification Time | 36–48 hours | 3.8 minutes |
| Manual Review Rate | 100% | 7% |
| Approval Accuracy | 94.5% | 98.8% |
| Drop-Off Rate (during KYC) | 37% | 11% |
| Compliance Officer Workload | ~6 hours/day | < 1 hour/day |
Why It Worked
- Make.com allowed for flexible, multi-layered logic
- Third-party KYC APIs enabled real-time validation with minimal cost
- OpenAI’s metadata parsing flagged false positives without hardcoding rules
- Airtable provided full auditability for regulators
- Slack alerts kept the human-in-the-loop only where needed
Looking to Automate Your Compliance or Verification Workflow?
We help fast-growing companies build secure, scalable automations for identity checks, document validation, and regulatory compliance — using Make.com, AI, and your tech stack.