As of 13 August 2026, AI can only partly predict when a deal will close.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Hand it to a person

   Second choice

   A person who owns the outcome does this end to end, worth it when the failure is dear.

3. Do it yourself

   The distant third

   A chat interface, power-user skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the CRM record for the opportunity and copy the current stage, stage-entry dates, expected close date, deal value and every recorded next step.
   2. Export or copy comparable won and lost deals, including their stage durations and the dates on which they closed or were marked lost.
   3. Open the latest meeting notes, emails and call transcripts, then paste only buyer commitments, deadlines, objections and outstanding approval steps into a working document.
   4. Paste the prepared deal information into the prompt and ask the model to produce a close-date range, assumptions, confidence level and slippage risks.
   5. Compare each stated fact in the response with the CRM record and meeting source, deleting any inference presented as a fact.
   6. Ask a sales colleague to challenge the range against procurement, legal, security and budget steps that the model cannot see.
   7. Record the checked range and its assumptions in the CRM, then update it when the buyer gives a new commitment or misses a date.

   Prompt

   Estimate when the following sales opportunity is likely to close. Use only the information supplied and do not invent buyer activity, dates or commitments.
   
   Company and opportunity: [COMPANY AND DEAL NAME]
   Deal value and currency: [VALUE]
   Product or service: [PRODUCT OR SERVICE]
   Current pipeline stage: [STAGE]
   Date entered current stage: [DATE]
   Expected close date in the CRM: [DATE]
   Sales cycle length for comparable won deals: [PASTE DATA OR SAY UNKNOWN]
   Previous stage durations for this deal: [PASTE DATA OR SAY UNKNOWN]
   Buyer contacts and roles: [PASTE DATA]
   Latest buyer commitments and their dates: [PASTE DATA]
   Open commercial, legal, procurement or security steps: [PASTE DATA]
   Last meaningful buyer activity: [DATE AND DESCRIPTION]
   Meeting notes and emails: [PASTE RELEVANT EXTRACTS]
   
   Return:
   1. A likely close-date range, not a false-precision single date.
   2. The most likely date within that range, clearly labelled as an estimate.
   3. A confidence level of low, medium or high, with reasons.
   4. The three strongest signals supporting the estimate.
   5. The three biggest reasons the deal could slip, including missing information.
   6. The next question I should ask the buyer to reduce uncertainty.
   7. A list of assumptions, separating facts from inferences.
   
   If the evidence is insufficient, say so and widen the range rather than guessing. Do not treat the CRM expected close date as independent evidence.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.