Proposal Evaluator and Risk Check

The Prompt

You are a critical evaluator. Assess a proposal against explicit criteria, surface hidden risks, and make the approval standard visible.

PROPOSAL TEXT:
{{proposal_text}}

DECISION CRITERIA:
{{decision_criteria}}

CONSTRAINTS:
{{constraints}}

ALTERNATIVES:
{{alternatives}}

Return exactly:
1) Framing check
   - what decision is actually being made
   - what inputs are still missing
2) Criteria and constraint split
   - hard constraints
   - scored criteria
3) Scorecard
   - criterion
   - score
   - rationale
   - confidence in the score
4) Risk register
   - risk
   - severity
   - likelihood
   - reversible or irreversible
5) Assumptions
   - assumptions that must be true for success
6) Strongest counter-case
   - best argument against approval
7) Recommendation
   - approve, revise, or reject
   - conditions
   - what would change the recommendation
8) 7-day validation plan
   - actions that reduce the biggest uncertainty first

Rules:
- Be explicit about unknowns, missing data, and weak evidence.
- Avoid false precision; explain confidence level instead of hiding uncertainty inside a score.
- Distinguish reversible and irreversible risks clearly.
- If the proposal violates a hard constraint, say that before scoring softer criteria.

When to Use

Use this before approving proposals, vendor decisions, policy shifts, or strategy changes where optimism bias can hide execution risk. It is especially useful when stakeholders already want to say yes and need a harder pre-commitment check.

Variables

Variable	Description	Example
proposal_text	The full proposal or summary under review	”New CRM migration proposal”
decision_criteria	How the decision should be judged	”Cost, timeline, adoption risk, legal impact”
constraints	Hard limits and non-negotiables	”No extra headcount this quarter”
alternatives	Real competing options, including doing less or waiting	”Delay migration, partial rollout, full rollout”

Tips & Variations

• Add historical outcomes or prior failed projects if you want better calibration on execution risk.
• Ask for a CFO, CTO, or legal lens only after the base evaluation is complete; otherwise the output can drift into role-play instead of analysis.
• If approval requires a pilot, ask for the pilot success criteria and stop conditions explicitly.
• For high-stakes decisions, include the status quo as one of the alternatives; that usually sharpens the recommendation.

Example Output

Recommendation: Revise before approval.

Condition: pilot with one region first and proceed only if support ticket volume stays below the agreed threshold.