What AI agents do you have running in production?
by•
There are so many new AI agent platforms ( @Wordware @Lindy @CrewAI @zapier and so on) that I'm finding myself curious how everyone is using them.
What AI agents are you using in production? What do they do? Are they working and reliable? What would make them better? Are they replacing roles? Augmenting existing ones?
And if you're not yet using agents in production, why not? What should these companies build that would make you adopt it quicker?
2.1K views
Replies
We have research, qualification, and CRM-sync agents running in production, but the surprising reliability boundary isn’t model accuracy—it’s evidence about external state. We don’t let a run call itself complete from an optimistic API response. Every action carries an idempotency key, source record, expected state, and a later reconciliation check. If a CRM write or public action can’t be read back at an exact record or URL, it stays pending. That gives us narrow autonomy without pretending ambiguous outcomes are success. Disclosure: I’m Ege, building Atlantic in SF. Curious whether anyone else measures agents by verified state transitions rather than task-completion claims.