As of 13 August 2026, AI can only partly detect fake contact details in your leads.

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. Export the leads from your CRM with a stable lead ID, name, company, email, phone number, website and source, removing fields that are not needed for this check.
   2. Gather known-good records, known-bad records, email bounce logs and any existing company or contact evidence so the model has examples to compare.
   3. Paste the prompt and the lead table into a chatbot, then save the returned status, evidence, confidence and next-check columns beside the original lead IDs.
   4. Ask the chatbot to produce a separate list of checks it could not perform, including reachability, domain ownership, current employment and whether a phone number is answered.
   5. Run the uncertain and likely-invalid records through a prospect enrichment product such as Apollo, and record the enrichment or verification evidence rather than copying its conclusion without checking.
   6. Open the company websites and relevant business listings for high-value records, compare the published contact details with the lead record, and use an appropriate business phone check where the decision matters.
   7. Update the CRM with retain, investigate or suppress temporarily, keeping the evidence and date of the decision instead of deleting uncertain leads.

   Prompt

   I need to screen a list of business leads for potentially fake or unreliable contact details. I will paste a table below.
   
   Use only the information in the table and the evidence I provide. Do not browse, invent verification results or state that an email address, phone number or person is real unless the evidence proves it.
   
   For each lead, return:
   - lead ID
   - status: likely valid, needs checking, or likely invalid
   - fields affected
   - specific evidence
   - confidence: low, medium or high
   - the next check needed
   
   Check for malformed email addresses, implausible domains, duplicate contact details, inconsistent names and companies, incomplete phone numbers, obvious placeholder values, and details that conflict with the supplied evidence. Do not treat a free email provider, an unusual name or an unfamiliar company as proof that a lead is fake. Separate technical invalidity from uncertainty about whether the person exists or works at the company.
   
   At the end, provide separate lists of leads to retain, investigate manually and suppress temporarily. Explain which checks require an email verification service, CRM evidence, a company website or direct contact. Do not recommend deleting a lead solely because it is uncertain.
   
   Known-good examples:
   [PASTE KNOWN-GOOD EXAMPLES]
   
   Known-bad examples or bounce records:
   [PASTE KNOWN-BAD EXAMPLES]
   
   Additional evidence and rules:
   [PASTE YOUR EVIDENCE AND BUSINESS RULES]
   
   Lead table:
   [PASTE LEAD ID, NAME, COMPANY, EMAIL, PHONE, WEBSITE AND RELEVANT FIELDS]

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