“The save failed” is a symptom. “The role changed, but the open tab kept showing an enabled Save button” is a much more useful starting point for investigation.
A fresh Ministry of Testing discussion raises a practical distinction: describing a failure is not the same as locating where expected and observed behavior first diverged.
Try this with one stateful browser bug:
• Record the starting state: user role, existing data, open tabs, and relevant prior actions.
• Capture the action sequence and a trace or network log, with sensitive data removed.
• Ask your AI assistant for the earliest evidenced divergence, the expected behavior, and the observation supporting its answer.
• Require it to separate facts from hypotheses and identify missing evidence rather than invent a cause.
• Replay the sequence from the same starting state to check whether the proposed explanation holds.
The first failing assertion is a clue, not necessarily the root cause. A stale session or earlier state change may have set up the failure several steps before the error appeared.
What would you include in a bug report so an AI assistant- or a teammate -can investigate rather than just summarize?