Schema and Entity: The Quiet Half of AI Visibility Most Firms Skip

Most conversations about AI visibility focus on content — write more, write better, get cited, build authority. All of that matters. But there's a second half of the equation that almost no firm pays attention to, and it's the half that decides whether the model actually associates your good content with your firm in the first place. That second half is schema and entity. It's deeply unglamorous. It is also the difference between being a real, named, retrievable firm to AI tools and being a vague collection of pages that happen to live at the same domain.

What "entity" means to a machine

When a search engine or AI tool reads the web, it isn't reading documents — it's building a picture of entities. An entity is a real-world thing the model can identify and disambiguate: a person, a firm, a place, a concept. The clearer the entity, the more confidently the model can attribute knowledge, citations, and recommendations to it. The fuzzier the entity, the more the model hedges, mis-attributes, or simply leaves your firm out of the answer.

For a professional services firm, the relevant entities are usually four:

• The firm itself, with its legal name, location, founding year, and specialties.
• The senior people, each with their role, credentials, and area of expertise.
• The services, each with its own definition, audience, and outcomes.
• The locations, if you operate in more than one city.

If a model can't cleanly distinguish between these — if your firm has three slightly different names across the web, or your founder is listed as both "Managing Partner" and "Senior Counsel" depending on the page — the model loses confidence and stops naming you.

What schema actually does

Schema is a small block of structured data, written in JSON-LD and embedded in your site, that tells machines exactly how to interpret your pages. It's not visible to human visitors. It is the most important thing on your site to AI tools and modern search engines, because it removes interpretation from the equation. Instead of guessing whether you're a law firm or a law librarian, the model reads "LegalService" and moves on.

The schemas that matter for almost every professional firm:

• Organization (or its more specific child types like LegalService, MedicalBusiness, FinancialService, ProfessionalService). Names the firm, its address, phone, founding date, parent company.
• Person for each senior professional. Name, role, credentials, sameAs links to LinkedIn and other authoritative profiles.
• Service for each named service offering. What it is, who it's for, where it's delivered.
• LocalBusiness if you serve clients in a specific geography. Address, hours, service area.
• FAQPage for any page that answers buyer questions. This is the schema most often surfaced inside AI answers.
• Article for blog posts and long-form pieces. Author, date, headline, key entities mentioned.

Each of these is a few lines of structured text. Most decent CMS plugins can generate them. The barrier isn't difficulty. It's that almost no firm's marketing team thinks of schema as their job.

The "sameAs" trick

The single highest-leverage piece of schema for AI visibility is the sameAs property on Person and Organization. It tells machines: this entity, on my site, is the same entity as the one over there. It's how you connect your bio page to your LinkedIn profile, your Wikipedia entry (if you have one), your Crunchbase page, your industry association profile, your podcast guest pages.

Once those connections are explicit, the model stops guessing. It can confidently merge what it knows about you across the web into a single entity record. That record is what gets cited when ChatGPT is asked for a recommendation. Without sameAs, the model often treats each profile as a separate, partial entity and dilutes your authority across all of them.

A short audit

If you want to know whether your firm has done the schema and entity work, open your site and view the source of your homepage. Search for the word "schema." If you find a script block of type application/ld+json containing Organization, an address, a name, and a sameAs array linking to your other profiles, you're doing the basics. If you find nothing, or only the WordPress-default WebPage schema, you're invisible at the entity layer regardless of how much content you publish.

Then do the same on your founder's bio page. You're looking for a Person schema with name, role, credentials, and sameAs links to LinkedIn, your firm bio, and any podcast or association profile. If it's missing, the model has no reliable way to associate your content with you.

What this fixes

Done well, schema and entity work fixes three specific problems most firms don't realise they have:

• Mis-attribution. The model knows about your work but credits it to a similarly named firm or person. Schema disambiguates.
• Under-attribution. The model knows about your specialties in pieces but doesn't connect them to a single named firm. Entity work consolidates.
• Invisibility in local AI answers. The model doesn't know where you operate or who you serve. LocalBusiness and Service schema fix this directly.

The work isn't glamorous. It rarely shows up in a strategy deck. But it is the structural backbone that makes every other piece of authority work — content, citations, press — actually accrue to your firm instead of evaporating into a vague impression the model can't pin down. Get this right once, maintain it annually, and your content does double the work for the next decade.