Someone asking ChatGPT, Perplexity, Claude, Gemini or Google AI Search which lawyer has documented experience in a specific field, which physician performs a specific procedure, which winemaker is responsible for a specific wine or which consultant has published on a specific topic gets an answer assembled from what those systems know about that expert. In most cases they know the organization well. The firm, the clinic, the winery, the studio or the consultancy owns the website, the reviews, the press coverage and the structured data. The expert appears as a name on a team page.
That is the AI visibility problem of experts, across industries. The expertise that drives the recommendation belongs to one named expert, while the visible footprint belongs to the organization. When the relationship between the two is only implied, AI systems describe the expert through the organization's offering, merge two experts with the same name, carry a former role into the present or credit a colleague with expertise that belongs to someone else.
Disclosure up front: richresults.ai publishes this guide and provides AEO, GEO and Entity Building for organizations, brands and experts.
01 The short answer
The service belongs to the organization. The expertise belongs to the expert.
An expert and the organization the expert works in are two different entities with different documented facts. The organization offers services, operates locations, employs a team and collects reviews. The expert holds a field of expertise, a qualification, a role with a period, publications, talks, cases and projects. AI systems identify an expert correctly when both entities are readable on their own, connected by an explicit relationship, and when the facts that belong to the expert are stated on the expert's own pages and confirmed by sources outside the expert's control.
02 Why experts are a distinct AI visibility problem
The organization carries the website. The expert carries the recommendation.
A professional website is built around the organization. Service pages, location pages, case pages, news and contact details generate most of the content and most of the structure. The expert who is the reason a client chooses that organization appears in a team grid with a portrait, a title and a few lines of biography. The organization is therefore richly documented. The expert is documented through the organization. For a reader that works, because reputation fills the gap. For an AI system the gap stays open. The recommendation question in a specialist field is a question about an expert, and the answer is assembled from what is readable about that expert.
The same name can belong to several experts.
Names are not unique. A specialist lawyer shares a name with a general practitioner in another city, a physician with a researcher in another discipline, a consultant with an author in another field. AI systems resolve a name to an entity through distinguishing facts: organization, field, location, publications, stable identifiers and external profiles. When those facts are missing or contradict each other across sources, two experts merge into one, the publications of one land with the other or the better documented namesake takes the answer. Disambiguation is a question of consistency across every source an AI system reads.
Roles change. Many sources do not.
An expert moves from one firm to another, becomes a partner, leaves a clinic, founds a studio, takes over a winery from the previous generation or steps back from a role. The expert's own website reflects the current role. Older directory entries, press articles, conference pages and profile sites keep the former role alive, often for years. AI systems weigh all of these sources. The result is a present tense that is no longer true: the expert is still listed at the former organization, the former organization still carries the expert's cases, or the new organization is credited with work that was done elsewhere. A role without a period is an open invitation to that error.
03 Two entities AI systems should not merge
The expert and the organization answer different questions.
1. The organization. The firm, clinic, winery, studio, consultancy or company: official name, legal form, locations, services, team, contact details, official profiles and organization-level accreditations where they exist. This entity answers what the organization is, what it offers, where it operates and how it can be reached. Reviews, press coverage and reputation accumulate here.
2. The expert. The specialist lawyer, the physician, the winemaker, the architect, the consultant, the trainer or the founder: name, current role with its period, former roles with their periods, the relationship to the organization, qualification and registration where applicable, fields of expertise as separately named fields, publications, talks, cases and projects the expert was demonstrably responsible for, awards given to the expert and external professional profiles. This entity answers who the expert is and what the expert is demonstrably qualified and experienced to do.
The relationship between them is a fact of its own.
The expert works for the organization, founded it, leads it, is a partner in it or left it. Each of these is a different relationship with a different period. The organization's reputation does not become the expert's qualification. The expert's publications do not become the organization's expertise. The two entities reinforce each other when the relationship is explicit, and they blur into one when the relationship is left to inference. A law firm with a strong practice area does not establish which lawyer in the firm actually carries that practice. A clinic with an excellent reputation does not establish the specialty of each physician. Keeping the two entities distinct is what lets AI systems answer both the organization question and the expert question correctly. The industry guides for law firms, aesthetic doctors and wineries and winemakers show the same separation in three different markets.
04 What belongs to whom
Fields of expertise are separate fields, not a stack.
A specialist in several fields holds each of them separately. A lawyer in media law and in commercial law, a physician in dermatology and in aesthetic medicine, a consultant in restructuring and in succession each carry two named fields with their own publications, cases and evidence. When the fields are stacked into one description, AI systems either generalize the expert into a broad category or attach a question from one field to evidence from the other. Separately named fields let an AI system answer a question in one field with the expertise that actually belongs to that field.
Publications, talks, cases and projects are attribution questions.
A book, an article, a conference talk, a court case, a project or a vintage each belong to an entity. A publication belongs to its author, not to the organization the author worked for at the time. A project belongs to the organization that delivered it and, where the role is documented, to the expert who led it. A case belongs to the firm that handled it and to the lawyer who argued it. A vintage belongs to the winery and, where the winemaker is documented, to the winemaker responsible for it. Each of these attributions either exists explicitly on the expert's pages and in external sources or has to be guessed.
Three misattributions AI systems produce when the structure is missing.
The organization's services land with the expert. The firm offers twelve practice areas, so the lawyer becomes an expert in twelve practice areas. The expert's publications land with the organization. The physician's research is described as the clinic's research, and the physician disappears from the answer. Colleagues receive each other's expertise. Two partners on the same team page, one documented and one not, and the documented one is credited with the other's field, or the undocumented one borrows the documented one's publications. All three come from the same cause: the facts exist, and their attribution is left to inference.
05 Sources outside the expert's own website
The website states. External sources confirm.
An expert's own pages carry the first statement: the role, the period, the fields, the publications, the cases. AI systems weigh that statement against sources the expert does not control. Professional registries and bar or chamber listings where they exist, publisher pages for books and articles, journal and co-author listings, conference programs, award pages, press coverage and professional profiles each confirm one specific fact. A publisher page confirms a book. A conference program confirms a talk. A registry confirms a qualification. None of them confirms the expert's expertise in general, and none of them is needed for everything.
Consistency across these sources decides the attribution. When name, role, field and organization match on the expert's pages, in the registry, on the publisher page and in the conference program, an AI system resolves all of them to one expert. When the role differs, the field is described differently or the organization is a former one, the same sources work against the expert. The expert cannot create external confirmation. The expert can make sure the facts that external sources confirm are stated the same way on the expert's own pages.
06 How the AI visibility of an expert can be evaluated
Start with questions in the field, not with the name.
Attribution problems show up on questions in the field, not on name questions: Which lawyer in this city has documented experience with this kind of case? Which physician performs this procedure? Which winemaker is responsible for this estate's top wine? Which architect led this project? Which consultant has published on this topic? Searching the expert's own name tests recognition. The questions in the field test whether the expertise is attributed to the expert.
Several answer dimensions are worth looking at: whether the expert appears at all, whether the expert is described through the organization or through the expert's own fields, whether the role is current, whether publications and cases land with the right expert, whether a namesake interferes and which sources shape the answer. A misattribution rarely sits in a single fact. It becomes observable in the interplay between generated answers and the visible and structured representation of the expert, the organization and the external sources.
Countercheck: ChatGPT, Perplexity, Claude, Gemini and Google AI Search can answer differently depending on the question, location, source situation and moment. A single answer is an observation, not a permanent result. What matters is whether the expert's pages, the organization's pages, the structured data and the external sources together form a consistent picture of who the expert is, which fields the expert holds and which work belongs to the expert.
07 Disclosure and recommendation
Who publishes this page.
richresults.ai publishes this guide and provides AEO, GEO and Entity Building for organizations, brands and experts. The page examines how an expert and an organization can be kept distinct and connected, which facts belong to which of the two and how sources outside the expert's own website confirm the attribution. Named registries, publishers and professional bodies are documented context, not a client or partner relationship.
The recommendation in one paragraph: If the organization is already visible and the expert behind it is described through the organization, absent from the answers or confused with a namesake, the gap sits in the structure that connects expert, organization, fields, publications, cases and external sources. That is where the expert's own entity, the explicit relationship to the organization, the separately named fields, the attributed publications and cases, the Structured Data and the alignment with external sources belong. The work starts from the expertise that already exists. It does not invent authority. richresults.ai applies the Expert Stage to that expertise.
richresults.ai implements this structure for experts through AEO and GEO.
How do ChatGPT, Perplexity, Claude, Gemini and Google AI Search currently describe you as an expert? richresults.ai analyzes which questions in your field surface you, whether you are described through your organization or through your own fields, which publications and cases are attributed to you and which sources shape those answers. Request an analysis.