Full case studies
Build Logs
What was built, for whom, what broke, and the result — 3–10 min reads.
My Voice Agent Kept Double-Booking Appointments — It Was a Race Condition, Not a Logic Bug
The agent checked availability before every booking and still double-booked slots. Here's why "check then write" isn't safe, and what actually fixed it.
My First n8n AI Agent Build Went Into an Infinite Loop — Here's What Actually Fixed It
My first AI Agent workflow in n8n just kept looping — same tool, same call, no exit. Here's what was actually happening under the hood, and what fixed it for good, tested one change at a time.
Fixing a Voice Booking Agent That Double-Booked Slots
Two callers, four minutes apart, both got the same appointment slot confirmed — a bug that only shows up under real concurrent load.
Why an AI Product-Fit Agent Got Compatibility Wrong
A compatibility question sounds simple until a customer's vague description hides which specific variant they actually have.
Why One AI Interview Coach Agent Failed and One Didn't
A technical coaching agent worked immediately. A behavioral one needed a persona and a framework before it stopped sounding like a textbook.
Fixing an AI Sales Agent That Promised the Wrong Refund
A RAG-based Shopify sales agent handled routine questions well — until it turned a vague FAQ line into an incorrect refund promise.
Which AI Model Hallucinates Least for Customer Support?
The cheapest model in a head-to-head customer support test was also the most likely to invent a policy that doesn't exist.
Building a WhatsApp AI Chatbot With Zero API Experience
No prior API experience, just a working WhatsApp agent by the end — including the wrong ID and the payload mistake that ate an hour.