find-customer

Find companies worth selling to. 

Paste your website. We infer your ideal customer profile, read candidate companies page by page, and return a ranked shortlist with a source behind every claim.

From your website to a qualified prospect list, in four steps.

One URL in. An ideal customer profile you approve, hundreds of companies discovered and screened, the best ones researched page by page, and a ranked shortlist out.

Step 01 · Understand

Your ideal customer profile, inferred from your own site.

Paste your URL. We read your services, case studies and customer logos, then show you the profile we inferred — industry, size, geography, the problem you solve — so you can correct it before any research starts.

Get your list
  • Editable before research
  • One sentence of context is enough
  • About a minute
acme-logistics.com/about
DHLMaerskXPOFlexport
reading /about4 of 12 pages
Profile
editable
Sells
AI automation for logistics ops
Industry
3PL · freight · last-mile
Size
100 – 1,000 employees
Buyer
VP Operations
find-customer AI

Read your services, case studies and customer logos. Correct anything before research starts.

Looks right — start research
You
Step 02 · Discover

Hundreds of candidates, not thousands of rows.

Search queries built from your ICP go out to several providers to find lookalike companies. Duplicates merge, obvious misses drop, and a cheap screening model cuts the pile to the companies actually worth reading.

Get your list
  • Multi-provider search
  • Cheap screen first
  • Politeness limits on every crawl
NFHL
Candidates
  • Discovered312
  • Duplicates−47
  • Off-profile−204
  • Worth reading61
Providers
Exa 142Google 118Bing 52
find-customer AI

Lookalike queries built from your ICP. Duplicates merged, obvious misses dropped.

cheap screen · 8s
Sales
Step 03 · Qualify

Every buying signal verified with a quote and a URL.

For each survivor we read the real pages and run your qualification checks. A check passes only when there is a quote from a page we fetched. Otherwise it is marked Not found — never guessed.

Get your list
  • Pass · Fail · Not found
  • Source URL per claim
  • Conflicts flagged
/careers/ops-coordinatorfetched 2m ago
Operations Coordinator (3 openings)

You’ll own daily dispatch across six terminals. Today the team runs on spreadsheets and phone calls, and we’re automating shipment status communication in 2026.

Verified
Checks
  • Hiring in operations
  • Stated automation goal
  • Recent funding
  • Uses a modern TMS
find-customer AI

A check passes only with a quote from a page we fetched. Gaps stay marked Not found.

never guessed
You
Step 04 · Rank

A target account list you'd actually put in your pipeline.

Scores are deterministic from check verdicts: ICP fit, problem fit, timing and evidence strength. Ten to twenty-five companies, ranked, each with why they fit, a reason to reach out now and the likely buyer.

Get your list
  • Deterministic scoring
  • Reason to reach out now
  • Likely buyer role
ICP fit × timing61 companies
NorthwindHarbor LinePinecrest
ICP fit →timing →
Shortlist
  1. 1. Northwind Freight94
  2. 2. Harbor Line88
  3. 3. Pinecrest81
find-customer AI

Northwind: 3 ops roles opened this quarter — reach out now.

VP Operations → likely buyer
Sales
You
Export

Everything you need to start, nothing to clean up.

Export the shortlist as CSV with scores, reasons and sources. Mark each company yes or no and the next run learns what a fit looks like for you.

Get your list
  • CSV export
  • Yes / No feedback
  • Runs stay inspectable
shortlist_run_8f21.csv
12 rows · scores · sources
CRM
12 accounts created
Sheets
synced · 9 columns
Sequence
3 drafts ready
Would you contact them?
Northwind Freight
YesNo
find-customer AI

9 yes, 3 no. The next run tightens the profile around what you kept.

learns from feedback
You
Evidence

Evidence for every account. What we found, where we found it and how sure we are.

See the checks
Verdict 01

Verified

A claim quoted from a page we fetched, with its URL. The only thing that passes a check.

Verdict 02

Not found

We looked and found nothing. The gap stays visible on the card instead of being filled with a guess.

Verdict 03

Conflicting

Two pages disagree — About says 80 people, a job post says 300. We show both and lower confidence.

Score

Confidence

Not a vibe. It counts verified checks, weighs conflicts and drops when the pages we needed didn't exist.

Qualification checks

The checks we run. Every one comes back answered, with a source.

None of these need a spreadsheet of your own. A check passes only on a quote from a page we fetched — otherwise it stays Not found.

Get your list
01

Hiring signals

Open roles in the team you sell into, pulled from careers pages and ATS boards — job titles, counts, and the language of the posts.

Ops Coordinator ×3/careers
02

Stated goals

Plans a company wrote down on its own site: annual letters, news posts, product pages. If they said it, we quote it.

“automate shipment status”/news/2026-plan
03

Size & stage

Employee counts from About pages and job posts, cross-checked so one stale number can't carry the verdict.

240 employees · 6 terminals/about
04

Geography

Where they actually operate — terminals, offices, service areas — not where the domain is registered.

US Midwest · 6 sites/locations
05

Tech & tooling

Stack mentions in job posts and integration pages, including the absence of one you'd expect.

“spreadsheets and phone calls”/careers/ops
06

Customer overlap

Who they serve, read off their own case studies and logo walls, matched against who you already sell to.

DTC brands · marketplaces/customers
Context

Steer the ICP with one sentence. Tight or wide, your call.

One sentence of context changes the whole search. Tell us who you actually sell to, what to prioritise or where to look — the ICP, the queries and the checks follow.

Get your list
1sentence
steers the run
4fields
you can edit
3queries
per provider
You typed

“We mainly sell to US logistics companies with 100–1,000 employees.”

Fewer candidates, higher pass rate. Good when you already know your lane.

ICP we inferred
Industry
3PL, freight brokerage, last-mile
Size
100–1,000 employees
Geo
United States
Problem
Manual dispatch & shipment comms
Queries we run
  • 3PL companies midwest hiring operations
  • freight brokerage 'customer updates' automation
  • last-mile delivery company 200 employees
Signals

The buying signals we look for. The ones that fire become your opener.

12 signals · 1 source each
  • 01
    Open roles
    Hiring in the team you sell into
    read from /careers
  • 02
    Funding round
    Fresh budget, dated and sourced
    read from press
  • 03
    New leadership
    A new VP usually means new tools
    read from /team
  • 04
    Product launch
    Scope changes create gaps
    read from /blog
  • 05
    Expansion
    New markets, offices or DCs
    read from /locations
  • 06
    Tech adoption
    Stack mentions in job posts
    read from /careers
  • 07
    Customer overlap
    They serve who your customers serve
    read from /customers
  • 08
    Stated goals
    Plans written on their own site
    read from /news
  • 09
    Partnerships
    Who they already integrate with
    read from /integrations
  • 10
    Content themes
    What they keep writing about
    read from /blog
  • 11
    Job post language
    “Spreadsheets” is a tell
    read from /careers
  • 12
    Geography
    Where they actually operate
    read from /about
How it behaves

Polite account research. Honest results.

4 rules · no exceptions

Company-level only

01

No personal emails, phone numbers or scraped profiles. You get the company, the reason and the buyer role — you decide how to reach out.

Get your list

Polite crawling

02

Rate-limited fetches, a seven-day page cache and no hammering of the same site twice. Politeness limits are settings, not afterthoughts.

Get your list

Nothing invented

03

A check only passes with a quote from a page we fetched. Gaps are Not found. Disagreements are flagged, not averaged.

Get your list

Inspectable runs

04

Every stage writes structured output to the run. Open it and see what was searched, read, screened and why.

Get your list
Who it’s for

Built for B2B teams selling something specific.

SaaS companies and the agencies that build for them. Each profile shows what we read on your site, where your buyers leave signals, and what a qualified card looks like.

DT

Developer tools

Your buyer documents their stack in job posts and engineering blogs. The easiest ICP to prospect with evidence.

Prospecting profile
SA

B2B SaaS

You know your ICP by heart. You need it applied to the web, with proof, not a CSV of 40,000 rows.

Prospecting profile
AI

AI SaaS

Ignore 'exploring AI'. Find the workflow you replace, done manually, at a volume that hurts.

Prospecting profile
AG

AI agencies

You automate a workflow for a vertical. You need the twenty companies still doing it by hand.

Prospecting profile
DS

Design agencies

Design gets bought at moments: a raise, a new CMO, a first designer hire. Timing-first prospecting.

Prospecting profile
DV

AI dev agencies

Your customer wants to ship AI and can't hire for it. That gap is public: long-open roles, legacy stacks.

Prospecting profile
FAQ

Questions before you run it. Short answers.

Where the companies come from, what we do with data, and what you get back.

Get your list

A description of the companies most likely to buy from you: industry, size, geography, the problem they have and who owns it. find-customer infers yours from your own website — services, case studies, customer logos — and lets you correct it before any prospecting starts.

Those are databases and enrichment tools: you filter millions of rows and get contacts back. find-customer is an AI research tool: it reads each company's actual web pages against your ICP and returns a short, ranked list where every claim links to the page it came from. Fewer companies, far more evidence — and no contact data, so nothing to clean up or comply with.

From search. We build queries from your ICP, run them across several providers, merge the results and screen them cheaply before any deep research. Nothing comes from a purchased database.

No. This version returns company-level results only: the company, why it fits, a reason to reach out now, the likely buyer role and the evidence. No emails, no phone numbers, no scraped profiles.

Five to fifteen minutes for a free run, depending on how many pages we need to read. You can leave the page — the run keeps going and lands at its own URL.

It is marked Not found. We never fill a gap with a guess, and a check only passes when we can quote a page we actually fetched. If two sources disagree, we show both and flag the conflict.

Yes. After we read your site we pause and show you the profile. Edit anything, add a sentence of context, then start the research.

A ranked shortlist of 8–25 companies. Each card has fit scores, a plain-English reason, a reason to reach out now, the likely buyer and every piece of evidence with its source URL. Plus a CSV.

Free for early teams

Give us your website. We’ll do the reading.

We run the research for you and send back a short list: companies that fit, a named buyer at each, and the source behind every claim.