We’ve been experimenting with AI-native development at Thumbtack, rethinking team structures and workflows for shipping AI-powered features. This blog shares what we’ve tried some things that worked, some that didn’t.

I’d love to hear how other teams are approaching this.

Are you rethinking team structure or iteration loops with AI?

What’s helped or hurt cross-functional collaboration?

How are you all approaching AI Evaluations & building user trust.