Paste your last closed-won deals. Claude finds what they share, turns it into exact filters and counts the lookalikes before you spend a record. It works in any MCP client connected to https://api.leadocean.io/mcp. Setup is on the MCP page.
The prompt
Here are my last 8 closed-won deals (industry, headcount, country, buyer title):
1. Financial Services, 320, US, Director of IT
2. Banking, 450, US, Director of Infrastructure
3. Financial Services, 280, US, Director of IT Operations
(paste the rest)
1. Find the values that repeat across at least 5 of the deals.
2. Call leadocean_list_enum_values for industry, employee_range, job_level and job_function
and map each repeated value to an exact string.
3. Call leadocean_count_leads with those filters.
4. Show me the filters, the count and which deals did not fit.
5. Do not call leadocean_search_leads until I say go.What it resolves to
| Filter | Value |
|---|---|
| industry | Financial Services, Banking |
| employeeRange | 201-500, 501-1000 |
| country | US |
| jobLevel | Director |
| jobFunction | Information Technology |
Enum values are exact strings. Step 2 looks them up so Claude never guesses "Finance" when the value is "Financial Services".
How many match
556 mailable people matched on 2 October 2026. Filters: industry Financial Services or Banking, employeeRange 201-500 or 501-1000, country US, jobLevel Director, jobFunction Information Technology.
Mailable means verified, catch_all_valid or catch_all, the count_leads default. The count is free. The deals above are placeholders, so your pattern and your count will differ.
What you get back
Search rows carry the fields below. These are placeholder values, not real records.
| Title | Company size | Industry | Country |
|---|---|---|---|
| Director of IT | 201-500 | Financial Services | US |
| Director, Infrastructure | 501-1000 | Banking | US |
| Director of IT Operations | 201-500 | Financial Services | US |
Variations
- Ask for the pattern in the top 3 deals by contract value only, then count again to see how much narrower the ICP gets.
- Add "list the deals that failed to match and propose a second ICP for them" to split a mixed customer base.
- Replace step 3 with a count per country, to see where your customers cluster against where the lookalikes are.
Records it spends
The enum lookups and the count are free. leadocean_search_leads costs 1 record per person returned, so a limit of 25 costs 25 records. Free gives you 1,000 records one-off with no card. Pro is $499 a month. See pricing.
To run this as a saved command, use the ICP from closed-won skill.
Related prompts
More prompts live in the prompts hub.
Turn your closed-won deals into a sized ICP, free
Free to start. No credit card. 1,000 records to spend whenever you like.
Get your free API key →