Someone Is Being Recommended Instead of You
When a buyer asks an assistant who to hire in your category, a name comes back. We tell you whose it is, why the model chose them, and what it takes to become the answer instead.
When a buyer asks an assistant who to hire in your category, a name comes back. We tell you whose it is, why the model chose them, and what it takes to become the answer instead.
In traditional search you could at least watch the battlefield. Rankings were observable, competitors' pages were public, and you could tell roughly where you stood by looking.
Assistant-led research happens somewhere you cannot see. A buyer asks a question, gets three names and a paragraph of reasoning, and forms a shortlist — all before visiting a single website. If you are not in that paragraph, you were never in the running, and nothing in your analytics will tell you it happened.
This is the most consequential blind spot in marketing measurement right now. Deals are lost at a stage that leaves no trace. The only way to see it is to run the questions yourself, systematically, and record what comes back.
That is what this layer does: it makes the invisible stage observable, so you can work on it.
Run on a schedule, reported in your review, and turned into a specific set of moves.
01
For a fixed set of buyer-intent questions, how often each business in your category is named — you included. This is the headline number, and unlike a ranking it maps directly to whether you are in the consideration set at all.
Tracked: per model, over time
02
What the models say about you when they do name you. Being described as the wrong kind of business, in the wrong market, or with services you dropped two years ago is more damaging than not appearing — and it is usually fixable within weeks.
Tracked: claim-level accuracy
03
Which pages, directories, publications and reviews the models lean on when answering for your category. This tells you precisely where authority is being read from — and therefore where to earn it.
Tracked: cited source map
04
When a rival enters or climbs the answer, we look at what changed: new coverage, a funding announcement, a content push, a directory position. Cause is more useful than the alert itself.
Tracked: change events with cause
05
For multi-location businesses, how the answer differs city by city. Models are often confident in one market and blank in the next, which points straight at where the entity work is missing.
Tracked: answer by geography
06
Which questions your category is actually being asked, and how that shifts. The winning query set in eighteen months will not be the one we start with, and the tracking has to move with it.
Tracked: evolving query set
The questions that come up when teams dig into this.
Conventional analysis looks at rankings, backlinks, and content gaps — all upstream proxies for visibility. This looks at the output: what a buyer is actually told when they ask.
Both are useful, and they often disagree. A competitor with a weaker backlink profile can be named more often because their entity is clearer and their content is more quotable. That disagreement is usually the most actionable finding in the report.
Monthly for most engagements, weekly in genuinely contested categories. More often than that produces noise — model outputs vary run to run, and you need enough samples for a trend to mean anything.
We always report the query set and the sample size so you can see how much weight a movement deserves.
That is one of the more common findings, and usually among the fastest to fix. Inaccurate descriptions almost always trace to stale or conflicting information in sources the models trust — an old directory listing, an outdated profile, a press mention describing a business you no longer are.
Correcting the sources and strengthening the accurate ones typically shifts the description within weeks, because you are not building new authority so much as removing contradictions.
Yes. Every report includes the actual answers, not just the scores. You should be able to read what a prospect would read, and we would rather you audit our interpretation than take the number on trust.
Often better than in crowded consumer ones. Niche categories have fewer entities competing for the answer, so entity clarity produces outsized returns.
The caveat is query volume: if almost nobody asks assistants about your category yet, tracking it tells you little today. The audit will tell you honestly which situation you are in.
The audit includes a competitor citation map showing exactly who is named in your place, and why.