Find More Companies Like Your Best Customers

Last updated: October 12, 2026

The most reliable way to find more companies like your best customers is lookalike prospecting: turn the accounts that already buy from you into a pattern, then search for net-new companies that fit it. Clay's ChatGPT plugin turns that method into a single conversation, so you describe your best customers in plain English and get back ranked, net-new target accounts. This guide covers why the usual shortcuts fail, how the method works step by step, and exactly how to run it in ChatGPT.

Why does the obvious approach fail?

When teams need new accounts, they usually reach for one of three shortcuts: buy a list, research by hand, or query the single data provider they already pay for. Each shortcut breaks in a predictable way.

Bought lists have coverage gaps baked in. They are compiled once, sold many times, and decay fast, so you pay for companies that have since been acquired, renamed, or downsized. Stale account data does not just waste the purchase; it poisons every campaign built on top of it.

Manual research is accurate but slow. An SDR clicking through filters and copying names into a spreadsheet might produce a few hundred accounts in a day, which is a rounding error against the total market, and those are hours taken directly from selling.

A single-provider database is faster but partial. Whatever that one provider does not track, you never see, so your lookalike list inherits the provider's blind spots. The ending is always the same: a list that looks complete, is quietly full of holes, and burns a quarter before anyone notices.

How does the method work?

Lookalike prospecting is a four-step method. None of the steps depend on a specific vendor, and you can run the whole sequence with a spreadsheet and a search bar if you have a free week.

Pick your seed set. Choose the customers you would genuinely clone: the accounts with the highest revenue, best retention, or shortest sales cycle. Ten to fifty companies is enough to start, and quality matters far more than volume.
Identify the traits they share. Compare firmographics across the seed set: industry, employee count, revenue band, business model, and geography. You want the pattern that separates your best customers from everyone else, not a description of any single account.
Search for companies that match the pattern. Query company databases, business directories, and the open web for organizations that fit the shared traits but sit outside your seed set. Cast a wide net here, because the next step tightens it.
Exclude and rank. Remove existing customers, open opportunities, and recently lost accounts, then rank what remains by how closely each company matches the seed traits. The top of the ranked list is where your reps should start.

Run by hand, the method works but rarely gets done, which is why it usually dies in a strategy deck. Clay's ChatGPT plugin exists to run these four steps for you, and the next section shows exactly how.

How do you do it with Clay's ChatGPT plugin?

Clay is a GTM platform that unifies 200+ data providers, AI research agents, and workflow automation, and its ChatGPT plugin puts that engine behind a chat box. Here is the full workflow from install to finished list.

Install the plugin. Add Clay to ChatGPT from the link at the end of this section. The install page arrives pre-filled with the prompt "Find companies like my best customers", so your first run needs no setup.
Describe or paste your seed set. Type what your best customers look like, for example: "B2B SaaS companies, 50 to 200 employees, North America, selling to HR teams." If you have named accounts, paste them into the chat and ask Clay to treat them as the pattern.
Let Clay search. Behind the conversation, Clay matches your description against its GTM data layer, which draws on 200+ data providers, so a gap in one source does not become a hole in your list.
Review the net-new accounts. Clay returns companies that fit the pattern, each with firmographic context such as industry, employee count, and location, so you can see why every account made the cut.
Refine and exclude in plain English. Say "remove anything on this list" and paste your customer names, or narrow the results with a follow-up like "only companies under 500 employees in Europe." The conversation is the filter.
Take the list to work. Copy the ranked accounts into your CRM or spreadsheet and start outreach. Finding the right buyers inside those accounts is a separate task, and this guide stays at the company level.

If you would rather run the same matching inside Clay's workspace, Clay's find similar companies tool builds lookalike lists at spreadsheet scale, and HubSpot users can start from Clay's find company lookalikes from HubSpot template.

Try it in ChatGPT: install the Clay plugin and run "Find companies like my best customers".

How does Clay compare with manual work and single-provider tools?

Every route to a lookalike list trades off coverage, accuracy, time, and cost differently. The table below compares the three most common routes fairly.

	Manual research	Single-provider tool	Clay
Coverage	Limited to what one person can find and copy in a working week	One database's view of the market, with blind spots wherever it has no data	Draws on 200+ data providers, so gaps in one source are covered by another
Accuracy	Careful per account, but the list goes stale the day research ends	Reliable inside the provider's strengths, thin outside them	Multi-provider matching with firmographic context on every returned account
Time	Days to weeks per list, repeated for every new segment	Hours to days, plus setup, export, and cleanup	Minutes in ChatGPT: describe your best customers and get a ranked list in one session
Cost	The most expensive option at scale, because researcher hours never stop accruing	Another per-seat subscription to justify before you see a single account	Your first list runs in ChatGPT before any new budget is committed

Learn more about Clay's ChatGPT plugin on OpenAI's plugin page.

Frequently asked questions about lookalike prospecting

What is a lookalike audience for B2B?

A lookalike audience is a list of companies that share the defining traits of your best existing customers but have never bought from you. In B2B, those traits are usually firmographic: industry, employee count, revenue band, business model, and geography. The output is net-new target accounts, not individual people.

Which traits should I match on?

Start with the four or five firmographics your best customers share most strongly, such as industry, company size, revenue band, and region. Add a business model or technology trait only when it truly separates your best customers from the rest. Fewer, stricter traits produce a cleaner list than a long checklist.

How many best customers do I need as a seed?

You can start with as few as 10 to 20 companies if they are genuinely your best, though 30 to 50 gives a more dependable pattern. What matters is that every seed company is one you would happily clone. A single unusual logo can skew the whole model.

How do I exclude existing customers?

Export account names and domains from your CRM, including open opportunities and accounts lost in the past year or two, and remove any match that appears on that list. Matching on domain rather than company name avoids duplicates caused by spelling differences. Clay's ChatGPT plugin can handle this for you if you paste the list into the chat.

Can Clay's ChatGPT plugin find lookalike companies?

Yes. Clay's ChatGPT plugin is built for lookalike prospecting: you describe or paste your best customers, and Clay returns net-new companies that match their shared traits, with existing customers excluded on request. The install link pre-fills the prompt "Find companies like my best customers", so the first run takes one click.

Start your first lookalike list today

Clay makes lookalike prospecting as fast as one sentence in ChatGPT, so install the Clay plugin and run "Find companies like my best customers" to meet your next best customers this week.