FROM BROKEN WEBHOOKS TO
100% DATA ACCURACY.
How Arsalan Faysal fixed a failing AI voice assistant pipeline – extracting complex call data and automating delivery with zero manual intervention.
Data Arriving,
But Extraction was a Mess
A business had built a Vapi AI voice assistant for intake calls. The goal was simple: extract child names, birth years, and call reasons into structured email summaries.
However, the Make.com integration was failing. The HTTP module was pulling stale or incorrect data, forcing the founder to spend hours manually cleaning up blank fields and incorrect records.
The Client Brief:
“Webhook is triggering correctly, but structured outputs are not being extracted. The module is pulling old call data and child names are constantly wrong.”
Fixing the Pipeline, Not the Band-Aid
I identified a caching issue caused by an improper payload structure and rebuilt the entire data flow in under 12 hours:
- 01. Webhook Payload Overhaul — Reconfigured Vapi to include unique call IDs and forced fresh data refresh per request.
- 02. Make.com Re-Architecture — Replaced broken HTTP modules with a custom webhook receiver that parses nested JSON correctly.
- 03. Dynamic Field Mapping — Used advanced iterators to cleanly parse child name, year, and intent without data bleed.
- 04. Automated HTML Delivery — Built a clean email summary template delivered to the founder within 2 seconds of call completion.
“Your module is pulling stale data because the payload isn’t forcing a refresh. I’ll rebuild the scenario with proper parsing and give you 100% accuracy within 48 hours.”
— Arsalan’s Initial Technical Diagnosis
The 12-Hour Fix
Diagnosis
Found caching issues in Vapi logs. Identified that nested JSON wasn't being flattened for Make.
Rebuild
Modified webhook URLs with unique parameters. Rebuilt the scenario: Webhook → Parse → Parse → Email.
Testing
Ran 10 simulated calls. Every field extracted correctly. Email delivered in under 2 seconds.
Handoff
100% accuracy achieved. Delivered a full video walkthrough and PDF documentation.
DATA INTEGRITY
“Arsalan fixed in 12 hours what my previous freelancer couldn't do in 2 weeks. Every single call now generates a perfect email summary. No more blank fields. Absolute game changer.”
— Alex (Founder)
🎯 The Takeaway for AI Voice Automation
Webhooks fail in subtle ways: caching, nested JSON, and missing identifiers. Success isn't "more code"—it's proper payload design. Flatten your JSON, use unique IDs, and test with real-world edge cases.
DON'T SCRAP IT.
FIX THE PIPELINE.
I debug and rebuild webhook + automation pipelines (Vapi, Make, Zapier, APIs) – fast. No fluff, just working data systems.
⚡ BOOK YOUR STRATEGY SESSION15 min. No BS. Just a blueprint to clean data flow.