There are a number of very tame applications that are normalizing like feeding logs from failed builds through the LLM.

Saves digging through the log, googling error messages etc.

Simple, low risk, eliminates engineer drudgery.

Absolutely, treat every story as an experiment, and monitor the results. It’s the only way to steer our products toward success!