Stuck on an AI-built app? Five checks to run before asking for another fix

A developer's guide argues that when a project built with tools like Cursor, Lovable, or Bolt breaks, asking the AI for yet another fix often makes things worse. Instead, the author recommends first reading the full console and server error messages, since they usually state the problem directly. Missing environment variables or API keys are flagged as the most common reason an app works locally but fails after deployment. The guide also suggests rolling back recently updated dependencies, and returning to the last working commit to request one small change at a time when the AI has added contradictory code. Finally, it advises reorganizing tangled code before adding patches, and seeking outside help after hours without progress or when real users and data are at risk.