Entity clarity. How a Portsmouth business defines itself to an AI engine, and why most do it wrong.

What entity clarity means

An AI engine that is asked "best independent dermatologist in Portsmouth NH" does not look up a directory. It assembles an answer from a long chain of inferences about which web pages describe which businesses doing which work in which place. Every step in that chain depends on the engine being able to identify the entity, the business itself, with confidence. If the engine is not sure whether two pages are about the same business, or whether a business operates at one location or two, or whether the dermatology practice is independent or part of a hospital system, the engine will hedge. Hedging looks like one of three things to the buyer: a vague answer, a wrong answer, or no mention of you at all.

Entity clarity is the set of signals across your website that let an engine answer six basic questions without ambiguity:

• What is the business. Category, in plain words. Not "your trusted partner." Not "premier solution provider." The actual category.
• Where is the business. One or more verifiable physical locations, stated identically wherever they appear.
• What does the business do. The actual services, named the way customers and the industry name them.
• Who runs the business. Principals, owners, or named clinicians, with credentials where applicable.
• How does a customer reach the business. Phone, email, hours, a contact path that resolves to the same entity.
• What proof does the business have. Reviews, licenses, association membership, third-party verification.

Six questions. If every page on your site answers them the same way, you have entity clarity. If your homepage says one thing and your About page says another and your Google Business listing says a third, you do not.

Why entity clarity comes first

We always start any GEO engagement with entity clarity work, even when the owner expected us to start somewhere else. The reason is mechanical. Every other GEO improvement compounds on top of an unambiguous entity. Schema markup, FAQ blocks, content rewrites, third-party citation cleanup, none of these pay back fully if the engine is still resolving two competing definitions of who you are. Fixing entity clarity first is what makes the downstream work measurable.

It is also the cheapest work in the program. The fixes are almost always textual. They almost never require a redesign. A capable owner with two hours and a clear checklist can close half the entity-clarity gap in an afternoon. The other half is structured data and third-party reconciliation, which is a longer job, but the structural conversation has already started.

The five places entity clarity goes wrong

1. The first paragraph does not say what the business is

The single most common defect. A homepage that opens with a tagline, a hero image with no caption, a video, or a marketing phrase, and then asks the reader to scroll for context. Humans tolerate this. AI engines do not. The first 200 words of your homepage are weighted disproportionately by every retrieval system we audit against. If those 200 words do not contain a plain-language statement of the form "Acme Plumbing is a residential and commercial plumbing service in Portsmouth, New Hampshire," the engine has to infer that sentence from indirect signals. Inference produces hedged answers.

The fix is one paragraph. Not a redesign. One paragraph, near the top of the homepage, that names the business, names the category, names the geography, and names the primary service line. Put the same sentence into the meta description while you are there. The number of Seacoast sites we audit that have a clean first paragraph and a vacant meta description is a category we did not expect to be as large as it is.

2. The business name on the homepage does not match the business name in the schema

Website builders (Squarespace, Wix, WordPress block themes, Webflow) often generate Schema.org JSON-LD automatically and stamp it with a default name pulled from the site title field, the theme template, or an old draft. The author then changes the visible brand on the homepage but forgets the schema. Engines see two names. They do not know which one to trust. They also pick up third-party citations of the wrong name and you end up with a quiet split between your brand and the engine's confidence in your brand.

This is a five-minute fix once you have noticed it, but most owners never look. Run your homepage through the Google Rich Results Test or any JSON-LD inspector and read the Organization or LocalBusiness block. If the name field does not match what is on your homepage and on your Google Business profile, fix it now. If you have a parent brand and a sub-brand (Acme Roofing under Acme Building, or a "doing business as" name), use the sameAs and alternateName properties to make that explicit instead of leaving the engine to guess.

3. The location is stated in three different ways across the site

We see this every week. The header says Portsmouth NH. The footer says Greater Seacoast. The Google Business listing says New Castle. The about page says serving New Hampshire and Southern Maine. Each statement is true in some sense. None of them are identical. An engine assembling a confident location signal has to pick one, and the one it picks is usually the one with the strongest third-party reinforcement, which is rarely the one the owner would have chosen.

The fix is the canonical-location rule: pick one statement of where the business is. Use it identically on the homepage, the footer, the contact page, the Schema.org markup, and the Google Business listing. The service area is a separate question that lives on a service-area page or in a clearly-labelled service-area block. The two are not the same thing, and engines benefit from the distinction.

4. The principal or named clinician is missing

For any category where the buyer cares who is doing the work, the named principal is a primary entity-clarity signal. Independent professional services (dermatology, dentistry, law, accounting, wealth management, design, architecture) live or die on this. The buyer is choosing a person as much as a business, and the engine knows that.

Sites that hide the principal behind a generic "our team" block under-perform across all eight engines in the categories where it matters. Sites that name the principal in the meta description (the strongest possible signal location), the homepage first paragraph, the Schema.org Person markup, and the About page are retrieved cleanly. Dr. Mendese, Mark Gianniny, Dan Hoefle. These are not just human-friendly names. They are entity-clarity anchors that AI engines weight heavily for professional-services queries.

5. The services list is written for marketing rather than for retrieval

We see services-pages that read like brochure copy. "Bringing your vision to life." "Solutions tailored to your needs." "Full-service support across every stage of your journey." These phrases tell the buyer nothing and tell the engine less. The engine needs the named services in the language the industry uses: "Mohs micrographic surgery." "Roof replacement and shingle repair." "Estate planning, probate, and elder law." "Whole-home generator installation."

The fix is mechanical. Replace each marketing phrase with the literal industry term. You can keep the marketing copy alongside for the human reader, but the named service must be on the page in plain language. If you sell ten things, list ten things. Engines retrieve named services for buyers searching by need. Buyers searching by need are the highest-intent traffic on the entire AI-search layer.

Schema.org markup is one entity-clarity tool, not the only one

Owners who have heard about GEO often arrive convinced that the answer is Schema.org markup. Schema is important, and a well-formed LocalBusiness or ProfessionalService block on the homepage moves you on Gemini, Copilot, AI Overviews, and Perplexity. But Schema by itself is not entity clarity. A page with perfect Schema and a vague first paragraph is not retrieved confidently, because Schema is a hint to the engine and the prose on the page is the evidence. Engines cross-check. Mismatches between Schema and prose are a red flag that the engine does not know how to resolve.

The right sequence is prose first, Schema second. Write a clean first paragraph that states what the business is, where, who runs it, and what it does. Then mirror those facts in the JSON-LD. The two should be saying the same thing, in different formats, on the same page.

The one-hour entity-clarity checklist

For any owner reading this who wants to do the work themselves before deciding whether to engage anyone, here is the checklist we use in the first hour of a paid engagement. It is the same checklist either way.

• Read your homepage like a stranger. Does the first paragraph state the business name, the category, the geography, and the primary service in one or two sentences? If not, write that paragraph.
• Read your meta description. Does it state the same four facts? If not, rewrite it. The meta description is the single most-weighted location signal in retrieval.
• Inspect the Schema.org JSON-LD on the homepage. Use Google's Rich Results Test or any JSON-LD validator. Confirm the name, address, telephone, and URL match the visible site exactly. If your site builder stamped a generic name in the schema, fix it.
• List every place your business name appears on the open web. Google Business, the chamber of commerce, your industry association, Yelp, your professional license listing, Facebook, Instagram. Check each one says the identical name. Reconcile the outliers.
• Find the principal. If your category is one where the buyer cares about who runs the business, ensure the principal's name appears in the first paragraph of the homepage, the meta description, and the About page, all spelled identically.
• List your named services. If the services page reads as marketing rather than industry-standard terms, rewrite each service line with the literal term the industry uses. Keep marketing copy alongside if you want, but the named service must be on the page.
• Pick one canonical location and one canonical service area. State the location identically on every page. State the service area separately, identically, on every page where it appears. Do not let the two collapse into each other.

Seven items. Two hours of work for most owners. The number of Seacoast businesses we audit that have completed all seven is in the low single digits per twenty. The number that have done none of them is a non-trivial share of the same twenty. Closing this gap is the cheapest GEO improvement in the field. Every downstream improvement multiplies its effect once entity clarity is in place.

What changes after the entity-clarity pass

Owners ask whether they will see results immediately. The honest answer is: on Gemini, Google AI Overviews, and Copilot, the cycle from a clean entity-clarity pass to a measurably better answer is roughly four to eight weeks, because those engines reindex regularly and weight structural signals quickly. On Claude and Perplexity the cycle is similar. On ChatGPT in search mode the cycle is faster than that because retrieval is live. On Meta AI and Grok the cycle depends on third-party signals more than on your own site, so entity clarity moves the needle there only when you also reconcile the open-web mentions.

The cleaner answer is what does not happen after the pass. Engines stop hedging. The vague-answer category disappears. The wrong-answer category disappears. The "not mentioned at all" category usually moves to "partially cited" with one pass, and to "consistently cited" after one or two follow-up passes that add structured data and FAQ blocks. The first pass is the one that unlocks the rest.

For the broader context on which engines you are even trying to be retrieved by, see the eight AI engines, in plain English. For the Portsmouth-specific field data behind these patterns, see how a Portsmouth business shows up in AI search (or does not). For the professional-services version see how customers find a dentist, lawyer, or CPA through AI search.