Competitive intelligence for the decisions AI makes about your market

The first question is about behaviour and is observed directly. The second is about judgement, and it is only observable by asking the systems doing the judging and reading what comes back. A competitor can change nothing at all and still start winning answers, because the material an assistant leans on shifted somewhere else entirely.

That is why this layer is rarely covered by an existing stack. Rank data describes pages, share-of-voice data describes coverage, and neither reports the outcome that a buyer now sees first: a short answer naming two or three businesses, with everyone else absent from it.

Treated properly, that outcome is an intelligence asset. It is measurable, it is repeatable, and unlike most competitive signals it maps directly onto a commercial consequence.

The unit of analysis here is the buyer question rather than the keyword. For each question a prospective customer would put to an assistant, there is a set of businesses the answer puts forward, and your business is named in it, absent from it, or one of several.

Read across the whole set, that becomes a landscape: which businesses hold which parts of the market's questions, where a category has a default answer, and where the answer is still contested. For a leadership team the pattern is usually more informative than any single answer, because it shows whether a weakness is broad or confined to one service line, one location or one kind of customer.

An assistant builds its answer out of what it can establish and what it appears to trust. So the meaningful comparison between you and the business being recommended is rarely about who publishes more. It is about what could be established about each of you, and which independent sources corroborated it.

This platform reports that difference as evidence: the stated reasons behind each recommendation, and the sources the assistants leaned on, including whether your business appears in those sources at all. Evidence differences are the part of the competitive picture you can act on directly, which is what makes them the right basis for a spending decision.

A single reading tells you where you stand. Intelligence needs the second reading, because the useful facts are directional: which answers you gained, which you lost, and whether the distance to the businesses AI names is closing or widening.

The measurement is versioned, and comparisons are made between comparable readings rather than between two differently shaped ones, so a movement is a movement rather than an artefact of a changed method. Movement of this kind belongs to the ongoing plans, because it needs more than one reading. The monthly plans are listed on the pricing page but are not open for self-serve purchase yet, and the recurring re-check is not switched on. What you can buy today is the free scan and the one-off report.

Each business the answers put forward gets a dossier: what it appears to do better in the eyes of the assistants, and how you compete with it on the strength of what those answers actually said. That is the difference between knowing a rival is winning and knowing the argument being made on their behalf.

The reading then ends where an intelligence product should end, in a decision. The fixes are ranked, the first three are called out, and each is tied to the evidence that argues for it, so the output is a sequence of moves with a reason attached rather than a list of observations for somebody else to interpret.

The plans are arranged as a ladder of questions. A one-off report answers where you stand. Monitoring is designed to show what changed, Growth to refresh the reasoning behind the change, and Professional is designed as the deepest self-serve competitive layer: the widest set of tracked competitors, refreshed competitive response and a re-ranked set of priorities. Enterprise is the same scope taken across an organisation, by contract.

Being plain about the current state matters more than the ladder. The monthly plans are listed on the pricing page but are not open for self-serve purchase yet, and the recurring re-check is not switched on. What you can buy today is the free scan and the one-off report.

This is not an attempt to replace every form of competitive intelligence software. It does not do advertising intelligence, social listening or app-store tracking. It is not a market-data terminal, and it is not a general website-change monitor: it does not keep watch over a competitor's pricing page, product launches or changelog, and it does not send real-time alerts the moment a rival changes their website.

There is no unlimited list of rivals either. Each plan tracks a set number of competitors, and the numbers are on the pricing page. Nor is there a promise about what any assistant will say, which nobody can offer.

What it does instead is measure one thing properly: the recommendation, and the evidence behind it. Teams that already run advertising, social and market-data tools keep them. This covers the layer those tools were never built to see.