Flipped Interaction
AI takes control of conversation, asking questions to gather sufficient context for complex tasks
What Is This Pattern?
The Flipped Interaction pattern instructs the AI to act as an expert diagnostician or requirements-gatherer, taking control of the conversation by asking the user questions one at a time until it has gathered sufficient context to perform a complex task. This solves the "low-context prompt" problem by having the AI guide the user through the information-gathering process.
How It Works
You instruct the AI to act as an expert in a specific domain (e.g., "Act as a senior SRE") and ask the user questions one by one, rather than expecting the user to provide all context upfront. The AI uses its domain expertise to determine what information is needed and asks for it systematically.
When To Use This Pattern
- Users don't know what information to provide
- Diagnostic or troubleshooting scenarios
- Requirements gathering for complex tasks
- Debugging workflows where step-by-step data collection is needed
- Expert consultations where the AI needs to guide the conversation
- Compliance audits or structured reviews
Example
Act as a senior SRE and ask me questions one by one (e.g., "kubectl describe pod", "kubectl logs") until you have enough info to diagnose the crash loop.Best Practices
- Use a strong persona pattern (expert role)
- Instruct the AI to ask one question at a time
- Have the AI wait for user responses before proceeding
- Use for diagnostic, troubleshooting, or requirements-gathering tasks
- Specify when the AI should stop asking and provide a solution
- Make it clear this is a multi-turn conversation
Common Mistakes to Avoid
- Not specifying the expert persona clearly
- Allowing the AI to ask all questions at once
- Not defining when to stop asking questions
- Using for simple tasks where direct instruction works better
- Not making it clear this requires multiple interactions