Amazon Connect AI Call Center Builder
There’s a workshop floating around that generates Amazon Connect contact flows and CloudFormation from a conversational UI powered by Claude Opus.
There’s a workshop floating around that generates Amazon Connect contact flows and CloudFormation from a conversational UI powered by Claude Opus.
A colleague pointed me at a Claude Code skill that strips AI-isms out of generated text. I needed it.
A client called me out — users were overwriting each other’s contact flows, and the pipeline got blamed.
I keep seeing the same pattern in Amazon Connect implementations: Lambda functions that return HTTP-style status codes — 200 for success, 400 for validation errors, 404 for not found, 500 for unhandled exceptions.
Client docs shouldn’t look like a README.
I’ve been Director of Engineering at CXBuilder for a few months now. Time for some honest reflection.
I’ve been promoted to Director of Engineering at CXBuilder!
A contact center handling 10,000 calls daily with just 8 seconds of sequential Lambda execution wastes 22+ hours of caller time every day.
Health check APIs with 200ms to 8-second response times at 200,000+ daily call volume? That’s a recipe for disaster.
I’ve joined CXBuilder as a Senior Architect!