← all playbooks

Find more companies like your best clients

A plain statement of what your best clients have in common, corrected by you, then a qualified list of companies that share it, each with the seed it came from and the reason it was judged similar. Hiring decision-makers with work email are added if you ask for them.

use when
you can name the five to ten clients you would clone if you could, and you want the companies that look like them
starts from
A list

The prompt

Paste it into Claude Code or the Claude desktop app with Hyreflow connected. The first line tells your agent to use Hyreflow, so it reads the play, asks before it spends anything, and hands the work back to you.

paste this into Claude
Use hyreflow and find me more companies like my best clients.

My best clients: <FIVE TO TEN COMPANY NAMES WITH WEBSITES>.
I picked them because <WHY THESE: MARGIN, REPEAT WORK, EASE>.

- Before you search for anything, tell me what these companies have in
  common: sector, size, location, what they sell and to whom. I will
  correct it. Do not expand until I have agreed the profile.
- Show me that each seed resolved to the right company, and point out any
  you were unsure about.
- Then expand from the seeds. Keep a column for the seed and the source
  each company came from, and the reason it was judged similar.
- Remove my existing clients and <EXCLUSIONS>. Keep to <GEOGRAPHY> and
  <SIZE_BAND>.
- Qualify what is left against the agreed profile, which is free. Only pay
  to check the borderline ones against their own website.

Expand from one seed first and tell me what the full run costs before you
do the rest. Contacts are a separate step: ask me before you find anyone.

Replace every <PLACEHOLDER> with your own detail. Everything else can stay as written.

What you need first

  • Five to ten of your best clients, each with a website or a LinkedIn company page
  • Anything that rules a company out: size, geography, sectors, existing clients, your clients' competitors
  • A Hyreflow workspace with credits

Tools it can reach for

The agent picks per step from what your workspace has. Nothing here is required by name.

What happens when you run it

Free steps are marked free. Anything that spends credits is marked, and the agent asks before the first paid run of any size.

  1. 1

    Resolve each seed to the right company

    credits

    Every client you name is matched to a company record and the match is shown back to you. A shared or ambiguous web address can resolve to a different business, and every lookalike after that would inherit the mistake.

  2. 2

    Say what the seeds have in common

    free

    Before anything is expanded, the agent writes down the pattern it sees: sector, size band, geography, what the companies sell, and the signs on a website that would prove a fit. You correct it, and the corrected version is what the run qualifies against.

  3. 3

    Expand from the seeds

    credits

    Several lookalike sources run from the same seeds, because each judges similarity differently. One grades matches in tiers from near-identical to broadly similar, one returns a score and a reason per match, one compares what the seed's own website says, and one takes the whole seed set at once.

  4. 4

    Merge and remove what you already have

    free

    Results collapse to one row per company by web domain, with the seed and the source kept on each row. Existing clients, your exclusions and anything outside your geography or size band are dropped using data already in hand.

  5. 5

    Qualify against the agreed profile

    free

    Every remaining company is judged against the profile you corrected, and the reason is recorded. It is reasoning over data already returned, so it costs nothing and runs across the whole list.

  6. 6

    Check the unclear ones on the web

    credits

    Where a record is too thin to call, the agent reads the company's own site and the public web for the proof signs in your profile and attaches what it found. This is the paid half of qualification, so it runs on the unclear cases only.

  7. 7

    Find the contacts, if you ask for them

    credits

    For the companies that pass, a company-scoped search finds the people who own hiring in your discipline. The work-email waterfall then runs in a set order and stops at the first hit, and addresses can be deliverability-checked.

What a run costs

Credits are spent per company the play actually works, and a lookup that finds nothing usually costs nothing. The two figures are the run where the first provider answers and the run where every lookup walks its full chain.

companiesif the first provider answersif every lookup walks the chain
25$12120 credits$21210 credits
100$47470 credits$81810 credits
500$2352350 credits$4054050 credits
1,000$4704700 credits$8108100 credits

Free before anything is charged

  • Say what the seeds have in common
  • Merge and remove what you already have
  • Qualify against the agreed profile

What moves the number

  • The channel. This play buys work email, because these are people you are approaching about work you want to win. A work address is the cheaper of the two to find.
  • Coverage on people search and work email. The chain stops at the first provider that answers, and only that provider bills.
  • How many companies survive the free filters. Everything dropped before the paid steps costs nothing.
  • Providers you connect with your own key. Those calls bill your account, not your credits.

An estimate, not a quote, priced at the volume credit rate. Your agent sizes the run against your own workspace and tells you what it will cost before it spends anything.

Your best clients are the brief

Ask an agency owner to describe the ideal client and you get a sector and a headcount. Ask which five clients they would clone and you get a better answer, because that list carries things nobody writes down: how those companies buy, how they decide, what kind of hiring they do. A lookalike search uses the companies themselves as the brief.

The risk is that it copies the wrong thing. Give it five clients and it will find whatever they share, whether or not that is the reason they are good clients. So the play has a checkpoint before any expansion: the agent states what it thinks the seeds have in common, and you correct it.

What you get back

First the profile, in plain words: what the seeds share and the signs that would prove a fit. Then a table of companies with name, website, size, sector and location, the seed and the source that surfaced each one, the reason it was judged similar, and a tier. Tier A was checked against the company's own site, with the evidence attached. Tier B fits on the record. Tier C is possible. Contacts appear only if you asked for them.

Variations worth knowing

Tighten or loosen the similarity. One source grades its matches from near-identical to broadly similar. Start tight, and widen when the near-identical pool runs out.

Split a mixed seed set. If your best clients fall into two distinct types, run each group on its own. A mixed set produces a blurred profile and a list that resembles neither.

Add timing afterwards. A lookalike is a fit, not a reason to call today. To be told when one of these companies starts hiring, use know when a target account opens a role in your discipline. For a whole market and its size, use build the target account list for a recruitment desk.

Where this goes wrong

A seed resolves to the wrong company. Shared web addresses and common names match the wrong business. The play shows each resolved seed before it expands, which is the cheap moment to catch it.

The pattern is an accident. Seeds that share a trait by coincidence produce a list built on that trait. The read-back step exists for this, and it only works if you answer it properly.

Lookalikes are often rivals. Decide before the run whether a client's competitor is a lead or a conflict, and say so.

Thin records at small companies. Small and young businesses carry little data, so similarity is weaker there and more of them land in tier C. Treat that tier as a reserve, not a call list.

Questions

Why does it describe my clients back to me before it searches?

Because a lookalike search copies whatever pattern it is given, accidents included. Five clients who happen to share a city will produce a list of companies in that city unless someone says the city is a coincidence. Reading the pattern back costs nothing and needs one reply from you. It is the moment to say that size is what matters and the sector can be ignored.

Why five to ten seeds, and not one or fifty?

One seed gives you that company's direct rivals, which is a different list. Fifty seeds blur into an average that resembles none of your clients. Five to ten is enough for a pattern to show and few enough for you to check that each one resolved correctly. One of the sources also needs at least five seeds to work from.

Why use several lookalike sources?

They disagree in useful ways. A database lookalike matches on company records. A website comparison matches on how a company describes itself, which catches businesses filed under the wrong industry. A company that turns up from two sources independently is a stronger lead than one that appears once, and the source column lets you see that.

How is this different from building a target account list?

A target account list starts from a description of a market and sizes all of it. This starts from proof: companies that already pay you. It is narrower, quicker to run, and better at finding companies a filter would miss. It will not tell you how big the market is.

Will it suggest my clients' competitors?

Often, yes. A company that closely resembles a client is frequently its rival, and whether that is a lead or a conflict is your call, not the agent's. Say in the prompt whether competitors of named clients are welcome or off limits. Existing clients are removed when you name them.