365,369 notices in.
4,204 opportunities out.

Everything published, from the people who published it

Notices come from the official publishers themselves, never scraped from a second-hand aggregator, and they arrive every day, overnight. You see a procurement when the market does — not a week later, when somebody has got round to re-typing it.

The part nobody thinks about until it bites is knowing when a collection was incomplete. A feed can hand back a partial answer that looks exactly like a quiet week, and a system that takes that at face value loses a month of contracts without anything appearing to go wrong. Ours refuses to record a collection as finished unless it can show it got everything.

• The whole life of a contract — from early notice to award
• History back to August 2024
• Gaps detected and retried rather than silently accepted

Every notice is read, not scanned for words

Keyword alerts are how most people do this, and they fail in both directions at once. They fire on “drone” in a catering contract for a venue that happens to ban them, and they miss a genuine opportunity that says “remotely piloted aircraft system” or “uncrewed surface vessel” and never uses the obvious word at all.

We use AI to read the full text of a notice — what is being bought, by whom, and what a supplier would actually have to do — and to judge it the way someone on your bid team would, if they had time to read every one. It returns four things:

A relevance score

How likely this is to be real work for an uncrewed systems supplier, so the strongest sits at the top of your list and the marginal is still there to look at if you want it.

A category

Services, hardware, counter-UAS, advisory, enabling, or a framework agreement — so a manufacturer and a survey operator can filter to quite different things.

Its reasoning

A sentence saying why it scored the way it did, shown on every result. A score you cannot argue with is a score you cannot rely on.

What it asks for

The airframe, sensors, quantities, operating conditions and certifications the tender states — from the notice and, where the buyer attaches them, its documents. Tell us what your company flies and holds, and each notice then says which of those requirements you meet. An explanation, not a second score.

Into your workspace, your inbox and your CRM

Results land in a list you can filter by score, country, stage, market, type of uncrewed system and what the tender asks for, and search in plain words. Your team marks each one — reviewing, bidding, submitted, won, lost, passed — with shared notes and, if you like, your own respond-by date. A view you use often can be named and saved, and new matches then arrive by email on the cadence you choose; a daily or weekly digest gathers what changed for you in one message.

Those marks drive everything else. Deadline reminders go to each person individually, so one colleague reading a reminder cannot silence it for the rest. A marked tender is watched from then on: if its deadline moves, its value is revised or a clarification is published, you are told what changed, in plain words. If you connect HubSpot, marking something creates a deal — and only a person marking it ever does, never a machine score, because an integration that dumps hundreds of auto-generated deals into a pipeline gets switched off in its first week. And a team dashboard gathers everything flagged — grouped by stage, named to whoever made the call, with a comment thread on each — so the decision and the discussion stay in one place.