As of 13 August 2026, AI can score your inbound website leads.

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 website form or CRM export and download the lead fields needed for qualification, removing fields that are not relevant to the sales decision.
   2. Write down your ideal customer profile, positive buying signals, disqualifiers, urgency rules and available follow-up routes in a plain document.
   3. Paste the qualification rules and the lead export into the prompt, replacing each bracketed section with your business information.
   4. Ask the chatbot to score the leads and return the evidence, missing information, confidence and next action for every row.
   5. Compare each drafted score with the stated criteria and the original form submission, correcting any score that relies on an invented fact or an unstated assumption.
   6. Send only the agreed high-priority leads to the relevant salesperson, and record the score, action and eventual outcome in your CRM.
   7. After enough outcomes have been recorded, compare converted and rejected leads with their original scores and amend the qualification rules where the pattern is consistently wrong.

   Prompt

   Score the inbound website leads below for sales follow-up.
   
   Business and ideal customer profile:
   [Describe what makes a good customer, including location, company type, size, problem, budget signals and buying timeframe.]
   
   Qualification rules:
   [Set out the criteria, weights if any, disqualifiers, minimum information needed and what counts as urgent.]
   
   Available routing options:
   [List the actions, such as contact within a stated period, ask a follow-up question, nurture, or disqualify.]
   
   Leads:
   [Paste a table or CSV export of the website submissions.]
   
   For each lead, return:
   1. A score using only the rules supplied above.
   2. The priority category and recommended next action.
   3. The exact supplied evidence supporting the result.
   4. Missing information that could change the result.
   5. A confidence label of high, medium or low, with a brief reason.
   
   Do not invent facts, infer sensitive characteristics, enrich the leads from outside sources, or treat missing information as negative evidence. Keep personal data to the minimum needed for this task. Put borderline cases in a separate section and explain which rule caused the uncertainty. Finish with a short list of any qualification rules that are unclear or contradictory.

   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.