We ran 48 of the questions a real buyer would ask an AI assistant, aimed at one long-established design firm's category. The firm came back zero times, absent, because there was nothing for the model to find and repeat.
People treat an AI's answer about a firm as a verdict, a judgment reached about who you are. What's actually happening is smaller and more mechanical: the engine retrieves documents, reads the passages inside them, and answers with whatever those passages say when enough of them agree. A paper trail does that work, built by whoever bothers to write it down.
That paper trail is where firms lose ground without knowing it. A Mid-Atlantic contractor we've worked with holds a 24-year state DBE/MBE certification, verified and current. It sits behind a CAPTCHA in a state registry. No AI crawler can get past that gate to read it, so the strongest trust signal this firm owns is invisible to every engine that might get asked about them. The fix isn't publishing more. It's getting that fact restated somewhere a crawler can actually reach: a case study, a press mention, a directory listing.
On your own site, padded and repeated claims about your own expertise mostly give an engine more material to reconcile. The lever that actually works is someone else saying the same thing about you: a client's own case study, a trade write-up, an award you won rather than entered. Ahrefs tracked roughly 75,000 brands and found that kind of off-site corroboration correlates with AI visibility far more strongly than backlinks ever did, though even Ahrefs is careful to call the finding a correlation.
None of this replaces ordinary search visibility. Every engine we've looked at still requires a page to be crawled and indexed before it can be cited at all. Ranking well is still the price of admission. It's just no longer the whole game once you're in the room.
Most agencies chasing AI citation measure the wrong things: keywords, clicks, rankings, because those are easy to count. The real target is harder to see, the trust and authority of the whole firm, built from actual history rather than a page score. A consumer startup with no track record has a genuinely hard version of this; there's nothing yet to draw on. A B2B firm with decades of projects and client relationships has the opposite problem. The record already exists. The work is finding where it hasn't been written down and filling it in. We call that Mining the Gaps.
The screenshot above is the operator console behind that work, running here on our own record as the test case. Every block in the wheel maps to one counted item, a project, a documented asset, a recognition signal, tracing a firm's path from expertise to completed work to industry recognition, not a decorative chart standing in for it. Built as part of our Interface Intelligence practice, its job is narrow: surface exactly where the record thins out, which is where the mining happens.
One more thing worth knowing: this isn't a fact you check once. An engine's answer regenerates every time someone asks, sensitive to phrasing, different session to session. That cuts both ways. A paper trail can change faster than a reputation ever could, but a single check tells you almost nothing. The only honest way to measure it is the way we found that design firm's zero: the same fixed set of buyer questions, run again every quarter, tracked as a rate.
We run that measurement for AEC and CRE clients now. Quarter over quarter, you see exactly what the engines currently find about your firm, and what's still missing. What we hand you is evidence, checked on a schedule.
