Home · Business · Sales · Pipeline & forecasting

PARTLY

As of 13 August 2026, AI can only partly score your inbound sales leads.

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 ita colleague

What the alternative costsThe supplied tool data gives no price for a human sales operations specialist.

If this goes wrong: your team follows low-value leads while a genuine prospect receives no timely follow-up.

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 your CRM or lead export and collect the fields available for each inbound lead, including lead ID, company, role, source, enquiry text, pages or assets viewed, stated need, location and previous contact history.
    2. Write down your target customer description, disqualifiers, qualification rules, positive buying signals and negative signals using the same terms your sales team uses.
    3. Add a sample of leads with known outcomes, such as progressed, rejected, unresponsive or won, and label the outcome without including information that was only learned after the decision.
    4. Paste the prompt and the criteria into a chatbot, then paste the lead data in a structured table with one row per lead.
    5. Compare the highest and lowest scores with the known outcomes and ask a sales colleague to inspect the evidence and the recommended action for a sample from each priority band.
    6. Adjust the criteria where the scores conflict with the known outcomes, then send only the agreed priority list and its reasons to the person responsible for lead follow-up.

    Prompt

    Score the inbound sales leads in the data below from 0 to 100 for likelihood of becoming a qualified sales opportunity. Use only the information provided and do not invent missing facts. Apply these criteria: [TARGET CUSTOMER DESCRIPTION], [MINIMUM QUALIFICATION RULES], [POSITIVE SIGNALS], [NEGATIVE SIGNALS], [DISQUALIFIERS], and [EXPECTED SALES PROCESS].

Return a table with these columns: lead ID, score, priority band, positive evidence, negative evidence, missing information, recommended next action, and confidence. Treat missing information as unknown, not as a negative signal. Separate fit from intent, and explain which evidence affected the score. Flag any lead that needs a human decision because the evidence is ambiguous or contradictory. Do not contact anyone, change any CRM record, or claim that a lead will convert.

Lead data:
[PASTE LEAD DATA HERE]

After the table, list the three scoring rules most likely to distort the results and the additional data that would test each one.

    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

  • AI cannot see unrecorded buying intent, internal politics or a prospect's changing priorities.
  • AI cannot decide whether a vague enquiry is strategically important when the evidence fits several interpretations.
  • AI cannot guarantee that enrichment, engagement data or CRM fields are current and accurate.
  • AI cannot take responsibility for missed revenue or for the sales team's follow-up choices.
  • AI cannot prove that a scoring rule will keep working when your market, offer or sales process changes.

What caps this at PARTLY: real time truth, judgement under ambiguity 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
Inputs1
Verification1
Liability1
Effort delta2
Total7 / 10

The methodology and its thresholds are published in full.

FAQ

Can AI score my inbound sales leads?
Partly. AI can rank leads from the data and rules you provide, but it cannot reliably infer intent that was never recorded or guarantee that the ranking reflects today's buying situation. Have a sales colleague test the scores against known outcomes before using them for routing.
What data does AI need to score leads?
Give it lead attributes, enquiry text, source, stated need, company and role information, recorded engagement and your qualification rules. Known outcomes from earlier leads make the result easier to test, but do not paste personal data that your organisation is not authorised to process.
Is AI lead scoring accurate?
It is only as reliable as the data, criteria and outcome labels behind it. Compare scores with leads whose outcomes are known, inspect the reasons rather than only the ranking, and reassess the rules when your offer or market changes.
Can AI automatically qualify and prioritise my leads?
It can apply a qualification rubric and produce a priority list, especially when your CRM data is structured. Keep human review for ambiguous, high-value or poorly documented leads, and do not let an untested score be the only reason a lead is ignored.

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.