One pass over 120M domains, not someone else's company list. You name the signals, and every company gets match or no match on each one, with the quote and the page behind every match. Write who you are looking for in plain language: the sector, the size, the place, the systems, the portal. We narrow the full database to the IAB verticals of your ICP, keep the live companies, screen every candidate against your criteria, and deliver a ranked list in which every signal carries a verbatim quote and the page it came from.
Company level only: no contacts and no named people. Pilot results within 1 business day, and the $100 is credited in full to a full run. Payment by PayPal or card.
Our research shows AI labs pay most for the records of real work, held in the systems below. When you start a pilot, tick the data types the companies must hold: we identify which of these systems each company runs and keep only companies that run a system for every type you need.
Every company comes with the systems we identified, shown as tags. Company level only: we show which systems a company runs, never what is inside them.
Twenty-six US industry lists already exist, and for many brokers they are the right place to begin. But a buyer's intake rules rarely stop at a sector. The moment your ICP adds a size band, names a system, asks for a client portal or rules out a state, a ready list gets you close and leaves the rest of the screening to you. That remaining work is what a custom list does.
Twenty-six US industry lists, from law firms and accounting firms to logistics and software, ranked by AI Data Value Score.
A screen of our full index against the ICP you write, with the evidence for every claim in the file.
A custom list is not a keyword search over somebody's company database. It starts from the whole active web, screens each candidate against the criteria you wrote, and shows its work for every company it keeps and every company it drops.
Every run starts from our full database of 120 million domains, 99.99%+ of the active internet, with domain history: the whole starting list, not someone else's company file. We narrow it to the IAB verticals of your ICP, for example the 93,020 US domains in IAB Legal Services for a law firm ICP, or the 160,054 in IAB Advertising and Marketing for agencies. So the starting pool is every domain in those verticals, not only the companies that happen to have a profile somewhere.
Before screening starts, every company domain is verified active and every company is checked for fit with its industry. A verified active domain resolves, is not expired, is not parked, and shows no error or placeholder page.
Your ICP becomes the filter, criterion by criterion. Every candidate is screened against it: is the team inside your size band, is the company in the US, does it name the systems you need, does it run a client portal, is it independent.
Every value comes with its quote and source page. A company whose site is silent on a must-have does not pass it.
Every signal we claim ships with a verbatim quote, the source page it came from and a confidence level. Companies that meet every must-have are ranked, strongest fits first.
Companies screened out go to an exclusion log with the reason, so you can see what was left behind and why.
Each stage narrows the pool for one reason only, and each reason is written down. This is the path every custom run takes, pilot or full.
The bars show the order of the stages, not their real sizes. How many domains survive each stage depends entirely on your ICP, which is why we do not quote a yield before your pilot has measured one.
No filter syntax and no field codes. One or two sentences make a complete brief. We split it into must-haves, which decide whether a company counts, nice-to-haves, which are reported and used for ranking, and exclusions, which remove a company whatever else it shows.
Robotics or physical AI team looking for real workplaces to record in? See site lists for physical AI data collection, with example briefs.
US accounting firms with 10+ staff that name QuickBooks or Xero
Reported alongside: a high client-confidential flag, since accounting work sits under engagement letters, and fit against Mode's published 10+ minimum for accounting firms.
Logistics companies of 30 to 200 people with a customer portal
Reported alongside: fit against micro1's published 30+ employees (its referral posting says 30 to 200), US first and primarily English.
Law firms with a help center and a client portal, outside California
Reported alongside: a high client-confidential flag for privileged matters, and fit against Mode's published 6+ minimum for law firms.
Every run reports a standard set of signals, for example team size, location, operating history and the systems a company runs, plus your custom signals. All of it is about the company, never about the people who work there.
From our category index. Your ICP can name one category or several, and can rule out close neighbors, such as software sold to accountants inside an accounting screen.
The team size a company states about itself, with the quote behind it. Use it as a size band in your must-haves, or only to rank.
Where the company says it is based, and whether it operates in the US. Use it to require the US, name states, or exclude one, as in "outside California".
Years online from our domain history. A founding year counts only where the company's own site states it, and then it comes with the quote, because a website can be younger or older than the company behind it.
Every company domain is verified active and every company is checked for fit with its industry. A verified active domain resolves, is not expired, is not parked, and shows no error or placeholder page. A domain that fails any of these never reaches your list.
Our 0 to 100 score for the data AI buyers want, with its grade, built from nine factor groups: history, scale, knowledge assets, operational systems, customer systems, organization, industry value, expertise and activity status.
The kinds of records a company probably keeps, such as support tickets, project files, case work, documentation or order history. Likely holdings from public signals, not an inventory of its systems.
Help desk, CRM, ATS, LMS, issue tracker, accounting and other business software the company says it uses for its own work.
A help center, docs site, knowledge base or published procedures. Written knowledge is among the records buyers name first, and it is usually the easiest kind to scope.
A login for clients or customers: a shipment tracking portal, a client file portal, a student or member area. A portal suggests structured, repeated records behind it.
Independent, PE or investor backed, or part of a group, as the company presents itself. Ownership decides who can sign, and many ICPs want only one of the three.
Whether a company's work involves client secrets or regulated records. The flag tells you, before you pitch anyone, that part of a company's records may be off limits.
Each company compared with the published rules of buyer programs: micro1's 30+ employees, US first and English, and Mode's 20+ US office staff, 10+ for accounting firms and 6+ for law firms.
Brokers and sourcing teams work toward a buyer's intake rules. A custom run can screen on those rules directly, or report them as columns next to your own criteria, so one file can serve more than one program.
| Rule | micro1 (as published) | Mode (as published) | Where it lands in your file |
|---|---|---|---|
| Company size | 30+ employees; the referral posting says 30 to 200 | 20+ full-time US office employees | Team size with its quote, and a fit reading per program |
| Industry exceptions | None listed | Accounting firms 10+; law firms 6+ | Industry from our category index, with the right minimum applied to each company |
| Location | US first, then other Western markets | US office employees; US-based teams the strongest fit | US or HQ location with its quote |
| Language | Primarily English | Not listed | The main language of the company's site |
| History and records | Mature operations | Several years of records | Years online, plus a stated founding year only with its quote |
| Company payout | "$100K-$2M+ for approved data packages" | "$100K-$5M" | Not a column: published ranges, not a price for any company |
Every value in a custom list has to be earned from the company's own website. When a run says a firm has "a team of 25", the sentence that says so sits in the next column, with the page it came from. When the site is silent, the cell says "not stated". It never holds a guessed number.
That rule costs some coverage, and we accept the cost. A firm with 40 people that never says so will read as "not stated" on team size, and it will not pass a team-size must-have. A broker who forwards a company to a buyer is spending credibility, and a list that guessed would be spending it for you.
Each value carries a confidence level, printed next to its quote.
The source column holds the full address of the page where the quote appears. Websites change, so the date of the reading travels with the file: a quote that has since moved still shows what the site said on the day it was read.
This is how one company looks in a custom list, laid out as a card in the same evidence style as our free check. The company, its domain and its quotes are invented for this example, and the domain uses the reserved .example ending. A real delivery carries the verbatim quotes and full source page addresses.
Our team of 18 CPAs and bookkeepers serves small businesses across central Ohio.
Visit our office on East Broad Street in Columbus.
Helping Columbus business owners since 2006.
You will close the books for clients in QuickBooks Online and Xero every month.
Log in to the client portal to upload documents securely.
Our help center walks you through uploading documents and approving your return.
An independent, partner-owned firm.
At least 18 people against 10+ for accounting firms, a US office, about 20 years of history.
The site suggests at least 18 people; micro1 publishes 30+ employees.
rank,domain,company,must_haves_met,team_size,team_size_conf,team_size_quote,team_size_source,systems,systems_conf,...,das_score,das_grade,status,fit_mode,fit_micro1
3,harborridge-cpa.example,Harbor Ridge Accounting,4/4,"at least 18",high,"Our team of 18 CPAs and bookkeepers serves small businesses across central Ohio.",harborridge-cpa.example/about/,"QuickBooks Online; Xero",high,...,78,B,active,likely,unlikelyExample, fictional company: the firm, domain, quotes and values are invented for illustration, and the .example domain is reserved and does not exist. An AI Data Value Score is an estimate from public signals, not a valuation or an offer, and not a sign that a company wants to sell data. Buyer rules as published, checked 7 October 2026.
A custom list is delivered as files your team can work right away, in a spreadsheet, a CRM import or your own scripts. Nothing to learn and nothing to log in to.
Every qualified company, strongest fits first, with every standard and custom signal, its confidence, the verbatim quote and the source page. One row per company, one set of columns per signal, so you can sort, filter and load it however you work.
Every company screened out, with the criterion it missed and the quote, or the silence, behind the decision. Use it to audit the screen, or to find near misses worth a second look, such as a firm one hire short of your size band.
The measured result of your pilot: how often companies in your universe meet your criteria, and what a full run projects to, with a likely range. It is the basis for every scope we quote after the pilot.
Domain, company name as its site gives it, signals, evidence and score. No contacts, no named people, no email addresses and no phone numbers. You reach the companies through your own network and your own channels.
Every reason is logged with its evidence. Silence is logged apart from a miss, because "not stated" and "stated, and outside your rules" lead to different next steps.
| Reason | What the log records | What you might do with it |
|---|---|---|
| Outside the size band | The stated team size and its quote, such as a team of 8 against a must-have of 10 or more | Set aside, or keep for a program with a lower minimum |
| Must-have not stated | Which criterion the site was silent on, and the pages that were read | Check the company yourself if it looks strong on everything else |
| Location excluded | The headquarters statement that placed the company outside your geography | Nothing, unless your geography changes |
| Different business on reading | The category matched, but the site describes another business, such as software sold to law firms rather than a law firm | Useful for a different ICP; the log names what the company actually does |
| Ownership excluded | Where your ICP asks for independent companies, the statement that the company is part of a group or investor backed | Look at the parent instead, which is often where a decision sits |
| Your own exclusion | Any custom exclusion you set, with the quote that triggered it | Confirm the rule removed what you meant it to remove |
Nothing in a pilot is prepared in advance. We set up the full run for your ICP, start it, and stop it the moment the first 25 companies qualify. You get those 25 in exactly the format of a full delivery, plus a measured projection of the rest.
On the pilot start page, describe the industry, size, location and systems you need, plus anything else a website could show. Mark what is a must-have and what is only nice to have, list any exclusions, and add your custom signals in plain language. Precise briefs make sharper pilots.
Your categories are pulled from our index, inactive domains are removed, and your criteria become the screen: the standard signals, your custom signals and the buyer rules you want reported. It is the same setup a full run uses, not a lighter version, which is what makes the pilot's numbers usable.
The screen runs on your universe with the same setup a full run uses, and stops as soon as 25 companies have met every must-have with evidence. We measure how often companies in your universe meet your criteria and project the full run, with a likely range.
Within 1 business day: the 25 qualified companies with every signal, confidence level, quote and source page, and the projection: how many qualified companies the full run projects to, with a likely range.
With the projection we quote several scopes: the full run, part of it, or a fixed number of companies. The $100 is credited in full to a full run. If you stop after the pilot, the 25 companies are yours to work.
We measure how often companies in your universe meet your criteria and project the full run, with a likely range. It is measured on the criteria you wrote, not borrowed from another client's run or from a sector average.
"The full run projects to about X, likely between L and H."
A projection is an estimate, so we report "about X" with the likely range, and each scope we quote is based on it. A tight ICP, such as law firms with both a help center and a client portal outside one state, projects lower than a broad one. The pilot shows you how much lower before you commit to anything larger than $100.
With the projection in hand, we quote several scopes up front. Whatever you choose arrives in the format your pilot companies came in.
Every qualified company in your universe, ranked, with full evidence.
The upper half of the ranking: the companies that fit your ICP best.
A smaller first batch for an opening outreach round, or for testing one buyer program.
Pick a company count that suits your budget, and the quote is for exactly that many.
Plans are quoted after the pilot; most clients choose a custom scope based on the projection. The reference plans below show how price follows the size of your universe and the number of signals.
Tell us the share of the projected yield or the number of companies you want, and which signals matter. We quote that scope, ranked strongest fits first. This is where the projection earns its keep: you pay for the slice you will actually work.
Screening a small niche and screening a broad category are not the same amount of work. The plans step up with the number of domains in your universe, and the pilot tells you that number for your ICP before you choose.
The standard set is built into every run. Each custom signal, such as "runs its own dispatch team rather than brokering loads", is new work: it has to be set up, tested on real sites and checked, so more of them move a run up the plans.
Payment by PayPal or card. Pilot results within 1 business day. Company lists, custom list pilots and runs, and all other one-time products are not refundable. API plans are the one exception: if you have not used your API plan yet, you can ask for a refund within 24 hours of purchase by writing to [email protected]. There is no free trial. The free tools let anyone test the method first: the AI Data Value Score demo and the check against buyer rules. To score your own domain lists by subscription, see the API plans and pricing; for one-time sector list files, see the list checkout. Last updated: 8 October 2026.
Custom lists are built for the people who find companies for AI data buyers, and for the teams inside data companies that need more of the right supply.
A referral pays only when a company signs and meets the program's conditions, so the work is finding companies that fit before you spend outreach on them. A custom list built to one program's rules puts those companies first, with the quotes you need to explain why each one fits.
As published, checked 7 October 2026. Each program's own terms decide who is paid, how much and when.
micro1, Mode, Troveo, Miro Advisory and others publish who they want, then need enough companies that fit. A custom run turns your written intake rules into a ranked pipeline.
Company level only. Your team makes every approach.
If your ICP is simply "US law firms" or "US software companies", a custom run repeats work that is already done. The 21 ready lists hold thousands of verified companies per sector, ranked by AI Data Value Score, and cost less.
To check a handful of companies you already know, the free check against buyer rules qualifies a target in about 20 seconds.
Pick the cheapest tool that answers your question.
The list is useful because it is disciplined. These are the lines every run keeps, in a pilot and in a full run.
Company level only. No names, no emails, no phone numbers and no personal profiles, anywhere in the file.
If a site does not state a fact, the cell says "not stated". We do not estimate a headcount or revenue that a company has not published.
A list shows companies that fit your ICP. It never claims a company wants to sell its data, and a score is not a valuation or an offer.
Every signal is about the company. If a requested signal would need facts about a person, we say so when we review the brief and suggest a company-level version.
We do not contact the companies on your list. How and whether to approach them is yours to decide, through your own channels.
A fit reading compares public facts with published rules. Buyers decide after their own review, and no buyer promises acceptance, so neither do we.
A ready list is one of our 21 US sector lists: thousands of verified companies per sector, ranked by AI Data Value Score, bought once as a file from $199 a list (the 3,000+ highest-scored companies), a snapshot with no updates. A custom list is built to your ICP. We screen our index of 120 million domains against your criteria, such as a size band, a named system, a client portal or a state to exclude, and every signal comes with a verbatim quote and its source page. If your ICP is one sector with no further rules, a ready list is faster and cheaper.
We configure the full run for your ICP, the same setup a paid run gets, and stop it when the first 25 companies qualify. You receive those 25 with every standard and custom signal, its confidence, the verbatim quote and the source page, plus a measured projection of the full run with a likely range. Results arrive within 1 business day. The $100 is credited in full to a full run, and the 25 companies are yours either way.
When it meets every must-have in your brief. Every value comes with its quote and source page. Silence is not a pass. A company whose site never states its team size cannot qualify on a team-size must-have, and in an exclusion log it appears with “not stated” as the reason. Nice-to-have signals are reported and used to rank, but they do not decide whether a company is on the list.
We measure how often companies in your universe meet your criteria and project the full run, with a likely range. The range can be wide, so we state the result as “about X” with that range and quote each scope from it. It is an estimate, not a guaranteed count.
No. Everything is company level: the domain, the company name as its site gives it, the signals, the evidence and the AI Data Value Score. There are no contacts, no named people, no email addresses and no phone numbers, in a pilot or in a full run. You approach companies through your own network and your own channels.
A standard set of signals on every run, for example team size, location, operating history and the systems a company runs, plus your custom signals: any criterion a website can show, written in plain language. Starter includes up to 2, Growth up to 5 and Proof up to 8. We do not screen on facts a website cannot show, such as revenue.
Pilot results arrive within 1 business day. Full runs are quoted after the pilot, because the size of your universe and the number of signals set both the price and the timing. Payment is by PayPal or card. Company lists, custom list pilots and runs, and all other one-time products are not refundable. API plans are the one exception: if you have not used your API plan yet, you can ask for a refund within 24 hours of purchase by writing to [email protected]. Reference plans run from $999 for Starter to from $25,000 for Platform, and most clients choose a custom scope based on the projection.
No. The pilot costs $100, and that is what keeps it a real run on your ICP rather than a canned sample. Two free tools let anyone test the method first: the AI Data Value Score demo at https://www.selldatatoai.com/data-asset-score/ scores a limited number of companies a day, and the check at https://www.selldatatoai.com/check-your-company/ reads a company’s public pages and compares them with published buyer rules, with quotes, in about 20 seconds.
$100, credited in full to a full run. Results within 1 business day, with every signal quoted and sourced, and a measured projection of how many more are waiting in your universe.