Decide

Analysis is done. Time to move from understanding to action.

Based on the synthesis above, guide the user through a decision. The goal is to stop analyzing and start acting, but with the right framing for the type of decision at hand.

Step 1: Decision Landscape

Before committing to a direction, present two to four viable paths that emerged from the synthesis. For each path:

• Path name: A short descriptive label
• What it looks like: One sentence describing the approach
• Primary strength: The single biggest reason to choose this path
• Primary risk: The single biggest concern with this path

This gives the user something concrete to react to before locking in a direction.

Step 2: Decision Type Detection

Assess which type of decision this is and adapt the output format accordingly:

Clear Choice

Use when: The evidence points strongly in one direction, tradeoffs are well understood, and the decision is relatively binary or technical.

Format:

• Decision: State what you will do. One sentence, direct, unhedged.
• Confidence: High / Medium / Low, with brief rationale
• Reversal triggers: Specific conditions that would reopen this decision
• Next action: The single next step to execute, something achievable today or tomorrow

Phased Approach

Use when: Significant uncertainty remains, the decision involves multiple stages, or a full commitment is premature.

Format:

• Initial move: What to do first and what it will reveal
• Evaluation milestone: A specific point where you will assess progress and decide whether to continue, pivot, or stop
• Conditional next phase: What happens if the initial move succeeds? What if it does not?
• Confidence: High / Medium / Low, with brief rationale
• Reversal triggers: Specific conditions that would reopen this decision
• Next action: The single next step to begin the initial move

Navigational

Use when: People, relationships, coaching, or interpersonal dynamics are central. The "decision" is more about an approach and principles than a single action.

Format:

• Approach: The overall strategy, framed as principles rather than rigid steps. "Lead with X, watch for Y, adjust based on Z."
• Opening move: The first conversation or action, with guidance on tone and framing
• Signals to watch for: What responses or reactions indicate the approach is working? What indicates it needs adjustment?
• Adjustment guidance: If the initial approach meets resistance, what is the fallback? How do you adapt without abandoning the core intent?
• Confidence: High / Medium / Low, with brief rationale
• Next action: The single next step to begin

Step 3: Synthesis Confidence Connection

If the synthesis confidence was Medium or Low, the recommended path must explicitly acknowledge this. Do not present a Medium confidence recommendation with the same certainty as a High confidence one. Options:

• Recommend a phased approach that builds confidence before full commitment
• Identify what specific information would move confidence higher
• Frame the decision as provisional, with a clear review point

Step 4: User Checkpoint

After presenting the landscape and recommended path, ask:

Does this direction match what you need, or should I reconsider from a different angle?

This is critical. The decide step should be a conversation, not a declaration. Wait for the user to confirm before treating the decision as final.

Decision Record

After the user confirms, provide a structured record for future reference:

## Decision Record
**Date**: [current date]
**Decision**: [one sentence]
**Type**: Clear Choice / Phased / Navigational
**Confidence**: High / Medium / Low
**Time horizon**: When this decision should be fully evaluated
**Review cadence**: How often to check whether reversal triggers have been hit
**Reversal triggers**: [list]
**Next action**: [specific step]

If you are not ready to decide, say so directly and identify what is blocking commitment. Then either:

• Run additional analysis on the blocking issue
• Accept that you are deciding under uncertainty, as most real decisions are

Do not use this as another round of analysis. Decide or explicitly defer.