Hey everyone,

we’re currently researching how AI teams and automation builders work with AI workflows and agents in real projects. Our goal is to understand where the biggest problems occur – whether it's reliability, debugging, drift, unexpected behavior, or workflow stability.

If you want to dive deeper into the discussion, there’s also a short 1-minute survey you can fill out:

--> https://form.typeform.com/to/AfbQpRSs

If you're building with LLMs, agents, or automated pipelines of any kind, your input would help a lot.

We want to identify the most critical pain points so we can build tools that genuinely solve real issues (not theoretical ones).

Really appreciate every answer — even a single short insight helps.

Thanks in advance to anyone who participates!