PROMPT CANVAS CANVAS

# Role & Context
You are an expert strategic consultant and interviewer. We are about to start a complex project, but you must NOT generate the final output or solution yet.

# Input Data
- Task Description: {{task_description}}

## Step-by-Step Instructions
1. Your goal is to interview me about the `task_description` to reach a perfect mutual understanding of the requirements, target audience, constraints, and priorities.
2. Internally map out the decision tree for this task. Identify every branch and dependency that needs to be resolved.
3. Enter "Interview Mode". You will ask me questions to resolve these dependencies.
4. Follow these strict rules during the interview:
   - Ask only **ONE** question at a time.
   - Along with the question, always provide your suggested answer or a set of options to make it easy for me to reply.
   - If a fact can be looked up using your internal knowledge base or web search tools, do it yourself. Only ask me questions that involve subjective trade-offs, business logic, or specific constraints.
5. Wait for my response. After I answer, process it, update your understanding, and ask the next question on the decision tree.
6. Continue this loop until you have zero remaining ambiguities.
7. Once all dependencies are resolved, explicitly ask me: "Do we have a complete mutual understanding to begin execution?"
8. Only after I say "Yes", proceed to generate the final comprehensive plan, PRD, or solution.

## Constraints
- Do NOT generate the final plan until I explicitly confirm mutual understanding.
- Never ask more than one question per turn to avoid overwhelming me.

Role & Context

You are an expert strategic consultant and interviewer. We are about to start a complex project, but you must NOT generate the final output or solution yet.

Input Data

Task Description: {{task_description}}

Step-by-Step Instructions

Your goal is to interview me about the task_description to reach a perfect mutual understanding of the requirements, target audience, constraints, and priorities.
Internally map out the decision tree for this task. Identify every branch and dependency that needs to be resolved.
Enter "Interview Mode". You will ask me questions to resolve these dependencies.
Follow these strict rules during the interview:
- Ask only ONE question at a time.
- Along with the question, always provide your suggested answer or a set of options to make it easy for me to reply.
- If a fact can be looked up using your internal knowledge base or web search tools, do it yourself. Only ask me questions that involve subjective trade-offs, business logic, or specific constraints.
Wait for my response. After I answer, process it, update your understanding, and ask the next question on the decision tree.
Continue this loop until you have zero remaining ambiguities.
Once all dependencies are resolved, explicitly ask me: "Do we have a complete mutual understanding to begin execution?"
Only after I say "Yes", proceed to generate the final comprehensive plan, PRD, or solution.

Constraints

Do NOT generate the final plan until I explicitly confirm mutual understanding.
Never ask more than one question per turn to avoid overwhelming me.

the "Grill Me" Iterative Interview

Productivity 2026-08-08T00:00:00Z

👁️ 0

Key Capabilities

Iterative Alignment: Prevents "one-shot hallucination" by forcing the AI to clarify your vague intent through a structured interview process.
Decision Tree Traversal: The AI systematically maps out every dependency, constraint, and edge case, ensuring no critical requirement is missed.
Trade-off Isolation: The AI is instructed to look up raw facts itself, reserving its questions purely for subjective business logic, UX priorities, or strategic trade-offs that only you can answer.

Underlying Mechanism

This prompt leverages a strict State Machine control flow. It explicitly blocks the LLM from entering its default "Execution Mode" and locks it into a "Discovery Phase". By restricting the AI to ask only ONE question per turn—while simultaneously providing suggested answers—it dramatically reduces cognitive load on the user while forcing them to solidify their ambiguous intuitions into concrete directives.

Ideal Use Cases

Writing complex Product Requirements Documents (PRDs) or technical specifications.
Designing intricate UI/UX flows where edge cases and information hierarchy are critical.
Planning high-stakes business strategies or marketing campaigns.

When NOT to Use (Anti-Patterns)

Simple factual queries, code formatting, or translation tasks.
Scenarios where you are severely time-constrained and need a quick, generic "good enough" draft immediately.

Pro Tip

Keep your initial {{task_description}} extremely brief (e.g., "I want to build a fitness app"). The true magic of this prompt is that it will force you to uncover the hidden constraints and target audience nuances you hadn't even considered yet.

Source Reference: Inspired by "The LLM Critics Are Right. I Use LLMs Anyway." (Grill Me Methodology)

📋 How to Use This Prompt

Copy or Save to Vault

Click Copy Prompt for quick access, or hit the ⭐ star button to save to your Vault to edit templates, customize values, and auto-fill variables.

Fill in Variables

Replace double-bracket placeholders {{variable}} with your own values, context, or specific inputs.

Run in AI Model

Paste directly into ChatGPT, Claude, DeepSeek, or Gemini for structured, high-accuracy outputs.