Food Truck Order System.

I built a technical proof-of-concept for an AI voice ordering system for food trucks at the end of 2023 and into early 2024. The market research looked promising on paper: some food trucks still rely heavily on phone orders, and cooking while answering the phone is a real operational problem. There was room for a tool that did the calls.

The technology wasn’t ready. The only competent option for realtime AI voice at the time was OpenAI, and the model wasn’t good enough even in clean conditions. Add in the actual conditions a food truck would call from — accents, poor cell connections, static, the messiness of real customers on real phones — and the experience got worse fast.

I wasn’t willing to sell a tool to small business owners that might cost them sales when it stumbled. So I stopped, and waited to see if the underlying models got there. They mostly have, now. I just no longer want to be the one building this particular thing.