Home · Business · Sales · Pipeline & forecasting

PARTLY

As of 13 August 2026, AI can only partly extract deal details from your emails.

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

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 ityou

What the alternative costsA purpose-built alternative is Apollo.io, described as a prospect database with AI outreach sequences and enrichment.

If this goes wrong: a wrong value, stage or next step enters your pipeline and distorts follow-up or forecasting.

What to actually do

  1. Use a tool built for this

    The route this page recommends

  2. Do it yourself

    Second choice

    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 email thread for the opportunity and include the full relevant conversation, including dates, senders, recipients, attachments converted to text, and the subject line.
    2. Remove unrelated personal information and any email thread that does not belong to the deal, while keeping the commercial facts and surrounding context.
    3. Paste the thread into the prompt and ask the model to extract only stated facts, quote its sources, and mark unclear or missing fields.
    4. Compare every populated field and supporting quote with the original emails, correcting names, dates, currency, value, stage and commitments.
    5. Resolve contradictions yourself with the sales owner and customer-facing record, rather than allowing the model to choose silently.
    6. Copy the checked fields into the CRM and record any unresolved questions as follow-up actions for the deal owner.

    Prompt

    Extract the deal details from the email thread below. Use only information explicitly stated in the emails, and do not fill gaps with guesses. Return a table with these fields: account, contacts and roles, customer need, product or service, quoted value and currency, pricing or commercial terms, current stage, decision makers, stated decision date, stated close date, next step, owner, competitors, procurement or legal requirements, risks, objections, customer commitments, and seller commitments. For every populated field, include the exact supporting quote and the email date. Mark each field as stated, unclear or missing. Separate facts from reasonable inferences, and put all inferences in a separate section with an explanation. Flag contradictions between emails and use the newest email only when the thread clearly resolves the contradiction. Do not create a probability, forecast, close date or CRM value unless it is explicitly stated. End with a short list of questions I should answer before updating the CRM. Here is the email thread:

[PASTE EMAIL THREAD HERE]

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

  3. Hand it to a person

    The distant third

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

What it gets wrong

  • It cannot access your inbox unless you provide the messages through an approved integration or export.
  • It cannot reliably tell whether a polite reply means genuine buying intent, routine procurement or a stalled deal.
  • It cannot invent a missing value, close date or stage without turning an absence of evidence into a false CRM fact.
  • It cannot know whether an email contains a binding customer commitment or an informal suggestion without the commercial context.
  • It cannot take responsibility for a pipeline record or forecast that is wrong.

What caps this at PARTLY: private data access, judgement under ambiguity and stakes of error.

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
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT read my sales emails and extract deal details?
Yes, if you paste the relevant messages or provide them through an approved workflow. It can extract stated facts, quote the source emails and format the result for your CRM, but it cannot safely decide what an ambiguous reply means.
Can AI update my CRM from my emails?
It can prepare CRM fields from an email thread, and some sales platforms provide workflows around prospect and outreach data. You should check each field against the source emails before an automated update, especially value, stage, close date and next step.
Is it safe to upload customer emails to AI?
Only use an approved tool and follow your employer's data, confidentiality and customer-consent rules. Remove unnecessary personal information and do not upload messages to a personal account if your organisation has not authorised it.
Can AI tell me which deals are likely to close from their emails?
It can summarise signals such as objections, commitments and stated dates, but those signals are not a dependable forecast by themselves. A salesperson still needs to confirm intent, commercial reality and the next action.

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.