Claude Prompts Description
This prompt tells the AI (as Grok) to take any chatbot script and, in a single ruthless pass, dissect its conversation logic: bluntly restating what the bot actually does, exposing where it assumes perfect users, pointing out missing branches, dead ends, pointless questions that don’t affect outcomes, loops that trap people, contradictory replies, and useless fallbacks; then stress-testing the flow against messy real behavior (vague answers, wrong intent, impatience, mid-conversation switches) to show exactly how and where the bot will frustrate users, escalate support, or cause drop-offs — and finally proposing specific, high-impact fixes to branching, decision points, context handling, and fallback design so the chatbot behaves like a robust decision system instead of a fragile happy-path script.
Reviews
There are no reviews yet.