We've been testing various agent frameworks for internal workflows, but reliability keeps being an issue. Between tool-calling errors, context drift, and unpredictable costs, it feels like we're still babysitting these systems. Has anyone found a setup that actually runs unattended? Curious if you're using LangChain
Are AI agents production-ready yet?
Replies
No comments yet — be the first to share your thoughts.