Technology in use
The vendor a company already runs, read from its own DNS and pages. High confidence only when two independent records agree.
Signalkeen reads what organisations publish about themselves: their pages, their DNS, their job boards, and the rules that bind them. When one approaches a decision you care about, you get a signal that names the page, quotes the sentence and gives the date.
Taking on first design partners.
“We will roll out a customer feedback programme across all three regions in the first half of 2027.”
An illustration of one signal. The domain is fictional; the shape is real.
The same signals answer different questions for different teams. Sales asks who to call. Product marketing asks who is switching. Strategy asks how big the market is. Partnerships asks who is building on what.
Which companies to write to this week, and the sentence they published that tells you why.
How it works →Organisations bound by a dated rule, with the date that applies to each and the rule’s own text behind it.
How it works →Companies running a rival’s product, and the moment that rival changes its price, plans or limits.
How it works →Run the same signals over your customer list: a new vendor’s script appears, or they start hiring for the job your product does.
How it works →Who uses which vendor, how many organisations face which date, by state, size or kind. Counts you can trace to pages.
How it works →Agencies and integrators who publish plans for, or hire around, the platform you build on.
How it works →Each reads one kind of public evidence, from sources allowed one by one. There is no general crawler, and nothing is bought from a data broker.
The vendor a company already runs, read from its own DNS and pages. High confidence only when two independent records agree.
What a company says it will do, in a sentence on its own site. The sentence must still be on the page, word for word.
A changed price, plan, limit or end-of-life notice at a vendor, applied to every company seen using that vendor.
A dated rule matched to an organisation’s country, kind and size, with the regulator’s own page and sentence.
Roles on a company’s own job board, read through the board’s public API. Never an aggregator, never LinkedIn.
Most intent data is a score built from people’s browsing. A Signalkeen signal is a fact about an organisation, with the page it came from. If the page changes, the saved copy still shows what it said. If the evidence is gone for good, the signal expires.
We run the signals and send one row per signal, as CSV and JSONL, best first. Open it in a spreadsheet, load it into your CRM, or hand it to your own agent. Nothing to install and nothing to log in to.
We build the list of organisations in a category from open sources, or run the same signals over a list of company domains you give us.
Signals read directly by your systems and by AI assistants, through a documented API and an MCP connector. Not built yet; design partners shape it.
Signalkeen runs today as a managed service, with two categories ready: survey and feedback software, and web accessibility for US state and local government. Every open signal has a saved copy of its evidence. With each design partner we keep a control group from the same list, written to the same way, so we can count together whether signalled companies reply more often. No customer logos, counts or quotes on this site, because we have none to show yet, and we would rather say so.
A category, a deadline, a competitor, a list of accounts. We will say plainly whether the signals can be built for it, what they would read, and what they would not.