AI search visibility

AI search visibility is the degree to which an AI system names a business when a buyer asks it a relevant question. It is the field Joe Ashta works in as a GEO / AEO consultant for law firms. This page explains the mechanism: how a system gets from a question to a named firm, why naming consistency matters, and why two visibility numbers for the same firm can disagree.

How a system arrives at a named firm

An AI answer that names a law firm is the end of a sequence. Each step depends on the one before it, and each responds to different evidence.

• Retrieval. The system searches for material relevant to the question and pulls a set of pages into its working context. This is grounding. A firm whose pages are never retrieved cannot be cited, and is rarely named. Retrieval needs relevance: the page has to be about the thing asked.
• Ranking. Among the retrieved pages, some are weighted more heavily than others. Here the signals resemble those of SEO: fit to the question, authority of the site, strength relative to competing pages.
• Recognition. The system identifies the firm as an entity: a specific organization with a name, a location and a practice, distinct from every other organization with a similar name. Recognition needs consistent, repeated statements of the same relationships, so that the pieces of evidence resolve to one firm rather than several candidates.
• Corroboration. The system finds the same relationships stated by sources the firm does not control: bar directories, legal publications, news coverage, professional bodies. Corroboration grows with the number, independence and authority of those sources.
• Dominant association. When the evidence is strong enough, the system prefers one firm for a given question over its competitors, and does so consistently. This is the state a firm is working toward, and it is the state in which the firm appears as a recommendation.

The sequence explains a pattern firms find puzzling. A firm can rank well and never be named, because it clears retrieval and ranking but fails recognition: the system has read its pages and still cannot say with confidence which firm they describe. A firm with a modest site can be named often, because third-party records have already done the recognition and corroboration for it.

Why naming consistency matters

Recognition and corroboration both depend on the same statements appearing in the same form across many places. The firm's name, its location, its practice areas, and the names and roles of its lawyers should read identically on the firm's site, on each lawyer's profile, on the firm's directory listings, and in the records third parties hold.

Variation is expensive. A firm listed under three name variants across its profiles, with two addresses and a practice description that changes from page to page, gives a system three weak candidates in place of one strong one. The system then has less reason to prefer that firm than a competitor whose records agree. Structured data helps because it states the identity in a form that needs no interpretation, and it helps most when the prose and the structured data say the same thing.

This site applies that rule to its own subject. The entity record states who Joe Ashta is once, in full, and every other page repeats the same form.

Why vendor-reported and measured numbers differ

Two kinds of visibility number exist, and firms often see one and assume it is the other.

A vendor-reported number comes from a data company, here Ahrefs, that crawls AI answers at scale and counts citations: the source links shown under answers. It is broad, inexpensive, third-party, and drawn from a sample of questions the firm did not choose. It shows which sites the platforms read.

A measured number comes from observing the platforms' own answers to the questions a firm's prospective clients ask, and recording whether the firm is named. It is narrower and costlier, and it is the only way to know what a specific buyer question returns. Its unit is the mention rate.

The two disagree because they count different events. A citation says a page was read. A mention says the firm was named. A firm can be cited often, on pages the platforms use for background, and rarely named as a firm to hire. Another can be named often on the strength of what other sources say about it, with few citations to its own site. The glossary keeps the two apart, and so does every number on this site. The BigLaw GEO Leaderboard is a vendor-reported instrument and is labelled as one.

What visibility means as an outcome

Visibility, as an outcome, is a firm being named in the answers its prospective clients receive, and named as a firm to consider. It is reported as a share of answers. The platforms do not agree with one another: a firm named by ChatGPT may be absent from Gemini, and a firm in Google's AI Overviews may be missing from AI Mode. A firm is visible when it is named consistently, across platforms, for the questions that lead to retainers.

The previous page, Answer Engine Optimization, covers the work that produces that outcome. The next, Law firm AI search, shows what the answers look like for a law firm and which kinds of practice are most exposed.