Inspiration
We came up with the idea for AI Market Tester because we’ve had many ideas ourselves, but often didn’t dare to act on them due to a lack of validation—or we would start doubting ourselves after just a small amount of skepticism and eventually abandon the idea altogether. Inspired by the recent wave of agentic AI apps, we thought about using agents to simulate how real people might think. This way, we can help early-stage founders—who don’t yet have users and can’t easily run real market research or A/B tests—quickly simulate user feedback with a synthetic persona panel and answer: What are the key issues with this idea? Which version performs better? Who is the best-fit audience? And who should we talk to next for real-world validation?
Lessons we learn
Database structure is most important thing to align when teammates are working on different components.
Challenge
At the beginning, we didn’t align on a unified database structure, so the front-end features didn’t map cleanly to the back-end functions. We ended up spending a lot of time coordinating and refactoring later to get everything consistent.
Built With
- nextjs
- supabase
- typescrip
- vercel
Log in or sign up for Devpost to join the conversation.