Find AI Prompts
Home›Sales›Deal Reviews
SalesPipeline & Forecasting

AI Prompts for Deal Reviews

A deal review inspects one opportunity closely: what do we actually know, from whom, and what would have to be true for this to close on the date in the CRM? Good reviews ask questions the rep cannot answer with adjectives. They produce a risk score that means the same thing across reps, and they end with coaching — the two things the rep will do differently — rather than a verdict.

These prompts structure the inspection, the risk scoring and the coaching. They work on top of the qualification framework and the mutual action plan; if neither exists for the deal, the review's first finding is that.

Before you use these

Have these ready to replace the highlighted [variables]:

The prompts

1. Inspect the deal with evidence questions

Best forA structured inspection that separates what is known from what is believed.
Inputs needed
  • Deal record and notes
  • Qualification evidence
  • Rep's summary
How to use itPaste the record and the rep's summary. The model generates the inspection questions, answers what the evidence supports, and lists what the rep must find out.
Expected outputInspection report: known / believed / unknown per area, the questions the rep could not answer, the close-date test, and the top three risks.
Act as a sales leader inspecting [opportunity] at [account] — stage [stage], amount [amount], close date [date].

Deal record and notes: [paste]
Qualification evidence: [per field with source]
Rep's summary: [their view]

1. For each area — pain and impact, economic buyer, champion strength, decision process and criteria, competition, timing and compelling event, paper process (legal/procurement/security), next step — classify what we know (with the evidence), what we believe (inference), and what we do not know.
2. Questions the rep cannot answer from the evidence, phrased as the questions to ask the buyer.
3. Close-date test: list the steps between now and signature with the buyer's process durations; state whether the date is arithmetically possible.
4. Contradictions between the rep's summary and the evidence.
5. Top three risks to this deal, each with the evidence and the action that would reduce it.
6. Stage and category recommendation, with the reason in one sentence.

Do not accept 'strong champion' or 'they love it' without the evidence. Confidence is not a field.

2. Score deal risk consistently

Best forA risk score that means the same thing for every rep's deals.
Inputs needed
  • Inspection results
  • Risk model or the loss patterns to build one
How to use itGive the inspection results and either your risk model or the patterns that predict losses. The model scores the deal and shows how each factor contributed.
Expected outputRisk score with factor contributions, comparison to the loss patterns, the factor that most drives the score, and the actions that would change it.
You are scoring deal risk for [opportunity] using a consistent model.

Inspection results: [known / believed / unknown per area]
Risk model: [factors and weights, or the loss patterns: e.g. no economic-buyer meeting by stage 3, single-threaded, close date slipped twice, no compelling event, competitor engaged first]
Deal attributes: [size relative to average, segment, source, rep tenure]

1. Score each risk factor present (0–3) with the evidence, and weight per the model; show the total and the band (low / medium / high / critical).
2. Pattern match: which loss patterns this deal currently matches, and the historical outcome for deals with that pattern (if data given).
3. The factor contributing most to the score and the specific action that would reduce it, with the date it could be done.
4. Factors that would raise the score if a pending event goes badly (e.g. security review).
5. The score's implication for forecast category, and the evidence that would justify a better category.
6. Calibration note: any factor where the model's weight seems wrong for this deal type, to feed back into the model.

Present the score as a small table. Keep the reasoning visible so two managers would reach the same score.

3. Turn the review into coaching

Best forTwo specific behaviors for the rep, not a list of everything wrong with the deal.
Inputs needed
  • Inspection and risk results
  • Rep's experience and patterns
  • Coaching style
How to use itThe model selects the highest-leverage coaching points, frames them as questions and actions, and writes the follow-up commitment.
Expected outputCoaching plan: the two behaviors, the question that opens each, the action and its date, the check-in, and what to acknowledge.
Act as a sales coach converting a deal review of [opportunity] into coaching for [rep].

Review results: [inspection, risk score, top risks]
Rep context: [tenure, patterns across their deals, what they do well]

1. Choose the two coaching points with the highest leverage on this deal and on the rep's pattern (not the longest list). For each: the observation with evidence, the question that opens the conversation without lecturing, the specific action on this deal, the date, and how the rep will know it worked.
2. Acknowledge: one thing done well on this deal, with the evidence — stated first.
3. The skill behind each coaching point (e.g. multi-threading, quantifying impact, process discovery) and one practice to build it beyond this deal.
4. The check-in: when, what evidence the rep brings.
5. What the manager owns (executive call, resource) so the plan is mutual.
6. The three-sentence version to say at the end of the review.

Keep the tone direct and respectful. No more than two points; a rep who hears eight fixes fixes none.

Related prompts

Logical next step

After this, most sales teams move on to Pipeline Review and Analysis.

Get the free B2B Sales AI Starter Kit → Nine of these prompts as a target → engage → qualify workflow with an intake worksheet, delivered by email. See what's inside

All Sales prompts · Search the full library