Telemetry and intent analytics for AI products using natural language

Hey all. About seven months ago, I was building a monetization/adtech tool for AI products. To support it, I built a small telemetry layer on the side mostly for internal usage.

I know it's a tired story for many here but .. I never thought it would happen to me. Our old users and most of our new users started to ignore the monetization tool and solely using our telemetry tool. Usage data, support conversations, churn patterns all pointed in the same direction. Six weeks ago we made our intent telemetry tool for the whole product.

What I built: A telemetry and intent intelligence layer designed specifically for LLM-based products. It captures what users are actually trying to accomplish, where conversations break down, which intents your model handles well vs. poorly, and how that connects to retention and activation. Any feedback would be great.

Documentation: https://docs.atheon-inc.com/

Demo: https://www.youtube.com/watch?v=54kusmA9a6c