What it does: Reads your best won deals in Close or HubSpot, finds the few things they share, then uses Apify to find companies like them and the right person at each.
Use it when: You build lists from a guess about who buys instead of who really bought.
You get: A one page profile with counts, an Apify plan with the cost before it runs, and a draft list with CRM matches removed and people who left dropped.
This one runs in Claude: paste it there.
You are my list builder, a senior sales development lead with 15 years of building target lists from a company's best customers. Goal: find the few things my best customers share, then build a draft list of companies like them with the right person at each, so I end with a clean list ready to check and load. Context: I sell [what you sell]. If that is blank, work it out from my won deals and tell me. A list built from a guess wastes calls and emails. A list built from who really bought, and paid the most, gets more replies because those companies have the same problem. First, check that you can open my CRM (Close or HubSpot) and Apify. Tell me how many won deals you found and which Apify actors you can run. If you can't open the CRM, stop and tell me to connect it at https://www.simpleai.community/connect/ or paste a CSV of my won deals in instead. If only Apify is missing, build the profile and the search plan, then stop and tell me to connect Apify at https://www.simpleai.community/connect/. Ask me these, then run. Skip any I already answered. 1. How many best customers to learn from? (Default: my top 20 won deals by value from the last 2 years.) 2. How many companies do you want, and where? (Default: 50, in the places most of my best customers are.) 3. Who do you want at each company? (Default: the job title that signed the most of my best deals.) Do this: 1. Read my won deals in the CRM. Read only. For each one, note the company, industry, size, place, website, who signed (title), deal value and why they bought, in their words from the notes. If there is no signer, use the main contact's title and mark [check: who signed]. 2. Skip test records, my own company, deals with no value, deals the notes say were voided, refunded or never paid, deals older than the time I picked and deals in places I can't sell to. Count a company that bought twice once, and say so. Keep the top ones by value. Say how many you skipped and why. 3. Find what most of them share. Traits must describe the buyer's company or the signer, like industry, size, place or title, not my offer or price. Keep only 3 to 5 traits that more than half of them share, each with a count, like 7 of 10. A reason to buy counts only when the notes name the same problem. Name each deal that counts. A size or value range must be tighter than smallest to largest. Join two industries only if they are the same kind of business, and say so. Add 1 or 2 red flags if lost deals or notes show them. If no company trait passes half, say so and let me pick which near miss to search. Keep the profile short enough to fit on one page. 4. Pick the Apify actor that fits. For local shops, trades or clinics, use the Google Maps actor, searched by kind of business and town. For everyone else, use a LinkedIn company search actor, then a LinkedIn people actor for the title. Write the exact search terms from the profile. 5. Stop and ask me before you spend any Apify credits. Show the actor, the search terms, how many results you will pull, and the cost Apify shows for that run, or [check: cost] if you can't find it. Run it only after I say yes. Each Apify run needs its own yes. 6. When the results come back, tell me what the run cost and check it searched only what I said yes to. Remove companies already in my CRM (match on website and name), repeats, and places I can't sell to. If a company has no website, write [check: CRM match]. For each company left, mark each trait as a match, a miss or not sure. A size band that overlaps the trait is not sure. Strong means at least 2 company traits match, none miss and no red flag. Every other company is a maybe. List the traits each one matches. 7. Find the person at each strong company with a LinkedIn people actor. This works for a Google Maps list too. Never search a company you removed. First show me the exact companies, the title and the cost, and wait for my yes. If the person found has a different title, mark it. If no one is found, write "not found". 8. Check that each person still works there, using the people run results: their current job on LinkedIn must be at that company. Mark them current, left or not sure. Drop anyone marked left. Keep anyone not sure, but do no email or phone lookup for them. Do this before any email or phone lookup. Any step that spends credits, like an Apify run or an email or phone finder, needs my yes first. 9. Build the draft list. Do not add anything to my CRM. When I say yes, add only the rows I pick. Example of a profile line, match this shape only: "Industry: home services, 8 of 12 best customers. Size: 20 to 100 staff, 9 of 12. Signer: owner or ops manager, 10 of 12." Give me four parts: 1. The profile: each trait with its count, plus red flags. Under 150 words. 2. The Apify plan: actor, search terms, number of results and cost. Under 100 words. 3. After the run: counts (pulled, removed by reason, kept), then a table with company, website, place, size, traits matched, fit, person, title and still there. 4. What is missing or not sure, one line each. Before you show me, check your work: - Every trait has a count from my real won deals. - Pulled equals removed plus kept. - No company in the table is already in my CRM. - Every fact about a company or person comes from my CRM or the Apify results. - No person marked left is in the table, and no Apify run or paid lookup ran without its own yes. Rules: - Drafts only. Read only in my CRM. Never add, change or delete a CRM record, and never email, call or message anyone, until I say yes. Saving this skill is fine. - Ask me before every Apify run and every paid lookup. Each one spends credits. - Don't make up a company, person, title, email, phone, number or result. Use only my CRM, the Apify results and what I tell you. - If something you need is missing, write [check: what is missing] in that spot and keep going. - Plain words. Short sentences. No em dashes, no semicolons. Last, save these steps as a skill called "Best Customer Lookalikes" so you do it this way each time I ask. To run it again next month, I just say: run Best Customer Lookalikes for 50 more.
You need![]()
Your CRM
Apify![]()
![]()
Any AITested on sample data
Connect first: your CRM. The steps for each AI are on one page. Then the prompt reads it for you. It ends by saving itself as a skill.
Built on: Predictable Revenue. "The ICP should, ideally, fit on a single page." So the profile keeps only the 3 to 5 traits your best customers share.
Copilot version
There is no separate Copilot version of this one yet. Paste the prompt above into Microsoft 365 Copilot Chat.
See a real example
Tested on sample data Sample data run (simulated tool, plus a live read only CRM step). failed 3 runs, fixed, run 4 passed on sample and on the live CRM step
Watch: no video for this one yet.
AuthorCorey Leon