Home · Business · Sales · Pipeline & forecasting

YES

As of 13 August 2026, AI can flag stalled sales opportunities.

Most people should hand this to a purpose-built tool.

Can you do it?

15 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededpower-user

Who has to check ita colleague

What the alternative costsThe available tool data gives no price for a sales operations alternative.

If this goes wrong: you waste time chasing a healthy deal or fail to revive a real opportunity because the CRM record hides the latest customer context.

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 your CRM pipeline and export the current opportunities with opportunity ID or name, owner, stage, value, last meaningful customer activity, next step, next-step due date, expected close date and recent notes.
    2. Remove unnecessary personal data from the export, retain the opportunity identifiers and dates needed for checking, and record the thresholds your team uses for inactivity and overdue next steps.
    3. Paste the redacted export and the thresholds into the prompt, then ask the model to return the requested table and separate definite data-based flags from items needing more information.
    4. Open each flagged opportunity in the CRM and compare the model's quoted evidence, activity date, stage history and next step with the live record.
    5. Ask the owner of each flagged opportunity whether there is recent customer context missing from the CRM, such as a verbal delay, procurement step or agreed meeting.
    6. Correct the shortlist using the CRM and owner answers, then send the confirmed actions to the relevant owners or record the agreed next steps in the CRM.

    Prompt

    You are helping a UK sales team review its pipeline. Analyse the opportunity data below and flag deals that appear stalled. Use the actual data only and do not invent activity, customer intent, dates or reasons.

Treat an opportunity as potentially stalled when there has been no recorded meaningful activity for [NUMBER] days, the recorded next step is overdue or missing, the opportunity has remained in the same stage for [NUMBER] days, or the expected close date has passed without an update. State which rule or rules triggered each flag. Do not treat an automatically logged email, unchanged value or repeated internal note as meaningful customer progress unless the data clearly says otherwise.

Return a table with these columns: opportunity name or ID, owner, stage, value if supplied, last meaningful activity date, next step and due date, expected close date, trigger for the flag, evidence quoted from the data, missing information, and recommended human follow-up. Separate high-priority flags from items needing more information. Do not recommend closing, downgrading or changing a forecast category without a human decision. End with five questions a sales manager should ask the opportunity owner before taking action.

CRM data:
[PASTE A REDACTED EXPORT OR TABLE HERE]

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

What it gets wrong

  • AI cannot know about customer conversations, procurement delays or internal politics that are absent from the CRM.
  • It treats incomplete or poorly maintained activity records as evidence unless you identify the gaps.
  • It cannot decide whether a quiet opportunity is strategically healthy, quietly lost or worth pursuing without the account owner's judgement.
  • It can prioritise the flags, but your team still has to contact customers, update the CRM and change the forecast.

Even on a YES, the friction has a name: judgement under ambiguity, private data access 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 AI tell me which sales opportunities are stalled?
Yes. Give it a current CRM export and explicit rules for inactivity, overdue next steps and unchanged stages, and it can produce a useful shortlist with evidence. Check every flag with the opportunity owner because the CRM may not contain the latest customer context.
What data does AI need to flag stalled deals?
Give it the opportunity ID, owner, stage, last meaningful customer activity, next step, due date, expected close date and recent notes. A redacted export is usually enough, but missing activity or inconsistent CRM updates will create unreliable flags.
Can AI update my CRM when a deal is stalled?
It can draft the changes or actions, but you should check them before updating the CRM. Do not let it close, downgrade or reforecast an opportunity without confirmation from the owner or sales manager.
How do I check whether an AI-stalled deal flag is right?
Compare the flag with the live opportunity record, stage history and recent customer activity. Then ask the opportunity owner whether any call, email, meeting or customer decision is missing from the CRM.

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.