AI made me code faster. It didn't make shipping a mobile app faster.

I've been using AI pretty heavily while building a production mobile app, and something surprised me.

Writing the code became much faster.

But the closer I got to actually shipping, the less coding felt like the bottleneck.

App Store review, subscriptions, background behavior, widgets, permissions, notifications, localization, testing on real devices, weird production-only bugs... none of that disappeared.

AI helped me get through parts of it faster, but “the feature works” and “this is ready for real users” still felt very far apart.

For people who have shipped an AI-assisted app to real users: what ended up taking way more time than you expected?