What we can see, what we infer, and what nobody can see from outside.

Things a company published about itself.

These come from live web research: the company’s own site, its careers page, its newsroom, and public coverage. They are facts with a date and a source, and the research carries the links it worked from.

• Funding events — a round, an amount, a date, as the company or the press announced it.✓confidence: verified — documented fact with a source
• Headcount and hiring — how many people, and what roles are open right now.✓confidence: verified — documented fact with a source
• What they build and who they sell to, in their own words.✓confidence: verified — documented fact with a source
• Sales and marketing stack, where it is detectable from their site.✓confidence: verified — documented fact with a source
• Recent news and announcements, with the coverage attached.✓confidence: verified — documented fact with a source
• Contact addresses published on their own pages, and whether a verification service could confirm each one.✓confidence: verified — documented fact with a source

What is not on that list matters as much as what is. We do not subscribe to a funding database, a filings feed, or a technographic provider. Everything above is research against public sources, done per account, at the point you asked for it.

Judgements drawn from shape and timing.

That a company with this shape, at this moment, probably has this need. Inference is honest work — it is what a good researcher does — but it is a probability, not a fact, and we mark it as one everywhere it appears.

• That the problem you solve is likely on their plate, given what they handle and at what scale.~confidence: inferred — probability drawn from company shape and timing, not a confirmed fact
• That budget is more available than usual, given a recent raise or a growth pattern.~confidence: inferred — probability drawn from company shape and timing, not a confirmed fact
• That a hiring pattern means the work has become someone’s job rather than an idea.~confidence: inferred — probability drawn from company shape and timing, not a confirmed fact
• Revenue band and company stage, where neither is published.~confidence: inferred — probability drawn from company shape and timing, not a confirmed fact

An inference the system can’t argue for doesn’t get made. If the reasoning can’t point at the evidence it came from, the account arrives without that signal rather than with a confident-sounding one.

Intent. Not by us, and not by anyone else either.

• Whether anyone there is currently looking for what you sell.—confidence: not observable — nobody can see this from outside the company, including us
• Whether they already have a supplier they are happy with.—confidence: not observable — nobody can see this from outside the company, including us
• Whether the budget survived the last board meeting.—confidence: not observable — nobody can see this from outside the company, including us
• Whether the person who would care has the standing to buy.—confidence: not observable — nobody can see this from outside the company, including us

No dataset contains this. Any tool that says otherwise is selling you a guess with more confidence than it earned — and the buyers most worth having are exactly the ones who notice.

Every no stays on the record.

When triage says no, the account is not deleted — it is marked discarded, with the reason stored. That is what makes the shortlist auditable: you can always see what was ruled out and why, not just what survived.

Does LeadAtomic have intent data?

No, and we don't buy it. Intent products infer interest from third-party browsing and content consumption, attributed to a company by IP or cookie. We treat that as weak evidence rather than a primary signal, and we would rather say a thing is not observable than dress a probability up as a fact.

So what makes an account worth contacting?

Fit and trigger. Fit is whether the company's shape matches what you sell to: what they do, who they sell to, how big they are, how they operate. A trigger is something documented and dated that changed recently — a funding round, a hiring pattern, a shift in what they're building. Fit without a trigger is a company that might buy some day. A trigger without fit is noise.

Isn't every prospecting tool guessing?

Yes. The difference is whether the guess is labelled. Every signal on a researched account is marked as something we confirmed or something we inferred, and the things nobody can confirm are shown as unobservable rather than quietly left out.