Home · Business · Sales · Pipeline & forecasting

YES

As of 13 August 2026, AI can estimate the value of a sales opportunity.

This still needs a person who signs their name to it.

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ita colleague

What the alternative costsThe supplied tool data does not state a price for a human sales operations review.

If this goes wrong: you overvalue a weak opportunity and make an inaccurate forecast or commit time and resources to the wrong deal.

What to actually do

  1. Hand it to a person

    The route this page recommends

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

  2. Use a tool built for this

    Second choice

  3. Do it yourself

    The distant third

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

    How to actually do it

    1. Open the opportunity record in your CRM and copy the deal amount, currency, sales stage, proposed close date, recorded probability, products, decision-maker information, competition, next step, and latest activity.
    2. Gather the relevant call transcript or meeting notes, qualification fields, customer requirements, objections, procurement steps, and any stated budget or timing evidence.
    3. Export or copy the historical win rate for comparable opportunities, together with your organisation's definitions for each stage and any rules for calculating forecast value.
    4. Paste the opportunity data and the historical rules into the prompt, replacing each bracketed slot and removing unrelated customer personal data.
    5. Ask the model to calculate the quoted value and expected value, then produce low, base, and high cases with every assumption shown.
    6. Compare the arithmetic with a calculator and compare every factual input with the CRM record, call notes, and current forecasting rules.
    7. Ask the opportunity owner and sales manager to challenge the probability, missing evidence, and proposed forecast category before entering the estimate into the CRM.

    Prompt

    Estimate the value of this sales opportunity using only the information I provide. Distinguish clearly between: 1) the quoted contract or deal value, 2) expected value calculated as deal value multiplied by probability of winning, and 3) any wider customer lifetime or expansion value. Do not invent figures, probabilities, customer facts, or historical benchmarks. If a probability is provided, use it but assess whether the evidence supports it. If no probability is provided, give a justified range instead of making up a point estimate. Show the arithmetic, list the evidence for and against the opportunity, identify missing information, and give a low, base, and high case where the inputs support that. State which assumptions are supplied facts and which are judgement calls. Finish with the three questions a salesperson should answer before changing the forecast. Use the currency and date format in the source data. Here are the opportunity details: [paste CRM fields, deal amount, stage, close date, probability, products, customer context, competition, next steps, qualification notes, and relevant call notes]. Here are the applicable historical win rates or forecasting rules: [paste them, or write 'none available'].

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

What it gets wrong

  • AI cannot know whether a prospect's stated interest is genuine without the relationship context and judgement of the salesperson.
  • It cannot supply a reliable probability when your CRM data is incomplete or historical win rates are not comparable.
  • It cannot detect every political, procurement, competitor, or budget factor that was not recorded in the source material.
  • It can make a neat expected-value calculation look more precise than the underlying evidence supports.
  • It does not own the forecast decision or the consequences of committing resources to the opportunity.

Even on a YES, the friction has a name: judgement under ambiguity, verification cost and context depth.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

AxisScore (0–2)
Output2
Inputs2
Verification1
Liability2
Effort delta2
Total9 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT calculate the expected value of a sales opportunity?
Yes. Give it the deal value and a probability of winning, and it can show the expected-value calculation and compare alternative scenarios. The probability still needs to come from defensible CRM evidence, historical data, or sales judgement.
What information does AI need to value a sales opportunity?
Provide the contract value, currency, stage, expected close date, probability, products, customer requirements, competition, next step, and relevant call notes. Historical win rates for comparable opportunities make the estimate more useful, but the model should label missing data rather than fill it in.
Can AI tell me whether a sales deal will close?
No, not reliably from a few CRM fields. It can weigh recorded evidence and produce a scenario range, but it cannot see unrecorded customer politics, a hidden competitor, or a change in budget.
Should I use an AI estimate in my sales forecast?
Use it as a challenge to the forecast, not as an automatic replacement for your sales process. Check the inputs and arithmetic, then have the opportunity owner and sales manager agree whether the probability and forecast category are defensible.

Nearby answers

All Pipeline & forecasting rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.