Entity Building: Making AI Understand Your Business Exists
An entity is a thing an AI system knows about — a business with properties like name, location, services, and reputation. Your entity is strong when many independent sources describe you consistently, and thin when you exist only on your own website. Thin entities get named far less often, regardless of website quality.
Two plumbing companies, same city, same size.
Company A has a beautiful website, a strong Google profile, and appears nowhere else on the internet.
Company B has a dated website and appears in fourteen places: Google, Yelp, Angi, BBB, the state license lookup, PHCC's member directory, two manufacturer dealer locators, the chamber of commerce, a local news piece about a water main break, a supplier's contractor list, two Reddit threads, and a high school sports sponsorship page.
Company B gets recommended by AI assistants and Company A does not.
That difference is entity strength, and it is the thing most local businesses have never worked on because it was never called anything.
What an entity actually is
Search engines used to match pages to queries. Modern systems maintain something closer to a knowledge model — a set of things they know about, each with properties and relationships.
Your business is one of those things, or it is not.
If it is, the system knows: this business exists, it is located here, it does these services, it serves these cities, it has these credentials, it has this reputation, it is currently operating.
Every one of those facts came from somewhere. Your website is one source. It is not a very trusted one, because every business says good things about itself. The confirming sources are what turn claims into facts.
Why thin entities lose
When an assistant decides whether to name your business, it is implicitly asking two questions: do I know this business is real and operating, and do I have something specific to say about it?
A business that exists only on its own website answers "maybe" to the first and "only what they claim about themselves" to the second. A business confirmed across fourteen independent sources answers both cleanly.
There is no penalty being applied. The assistant simply reaches for what it is confident about.
Diagnosing yours
Three checks, twenty minutes.
1. Count your mentions. Search "Your Business Name" in quotes. Count results that are not your own website or social accounts. Under eight is thin. Under three is why you get no AI mentions.
2. Ask an assistant directly. "Tell me about [your business] in [your city]." Read the answer carefully. Vagueness means a thin entity. Errors mean bad facts somewhere public. "I don't have information about that business" means no entity at all.
3. Check coherence. Do the sources that mention you agree on your name, phone, address, and what you do? Contradictions are worse than absences.
Building it
Roughly in order of effort-to-value.
Tier 1 — do this month
The major directories. Google, Bing Places, Apple Business Connect, Yelp, Facebook, BBB, Angi, Thumbtack, Nextdoor. Claimed, complete, and consistent.
State license lookup. Make sure your license is current and the registered business name matches what you use publicly. This is a government-run, independently verifiable source, and it is one of the strongest single signals you can have.
Trade associations. PHCC, ACCA, NARI, NAHB, ISA, NPMA, ASE, your state and local chapters. Membership usually comes with a public directory listing. Most businesses that pay dues never confirm their listing is correct.
Manufacturer and supplier directories. Dealer locators, certified installer programs, pro networks. Carrier, Trane, GAF, Rinnai, Pentair, whichever apply. These are free with relationships you already have and almost nobody claims them.
Chamber of commerce. Modest cost, public listing, local credibility.
Tier 2 — this quarter
Local press. Not a press release blast. Actual local interest: a storm response, a hiring milestone, a charity job, an apprenticeship program, a genuinely newsworthy repair. Local outlets are chronically short of stories and will often take a good one.
Local roundups. Find the "best plumbers in [city]" articles that assistants already cite — run the query in Perplexity and read the citations. Many are maintained by small local publishers who will consider additions.
Sponsorships. Youth sports, school programs, community events. Each produces a page with your name on it, on a domain that is not yours.
Supplier case studies. Your distributors and manufacturers publish contractor spotlights and are usually short of volunteers.
Community involvement. Chamber committees, trade school advisory boards, apprenticeship programs. Each has a public page.
Tier 3 — ongoing
Genuine community presence. Reddit, Nextdoor, and local Facebook groups. Participate honestly — answer questions in your area of expertise, disclose that you are in the trade, do not pitch. Over time people mention you, and community discussion is weighted heavily in AI answers.
Never astroturf. Fake accounts are against platform rules, reliably detected, and the reputational cost of being caught vastly exceeds the benefit.
Reviews with substance. Covered elsewhere, but they are entity signal too — recurring specific praise is a property of your entity.
Consistency maintenance. Every new listing must match canonical. One sloppy entry undoes several good ones.
Connecting the sources
Two mechanisms tell systems that all these listings are one business.
sameAs in your schema. List every profile URL:
"sameAs": [
"https://www.google.com/maps/place/?q=place_id:ChIJxxxxxxxx",
"https://www.yelp.com/biz/carter-plumbing-tampa",
"https://www.bbb.org/us/fl/tampa/profile/plumber/carter-plumbing-xxxxx",
"https://phccweb.org/directory/carter-plumbing",
"https://www.rinnai.us/find-a-pro/carter-plumbing"
]
Exact consistency. Same name, same phone, same address format everywhere. This is what lets a system merge the sources into one entity rather than treating them as several similar businesses.
What does not work
Directory blasting. Automated submission to 300 low-quality directories was a 2012 tactic. Modern systems weight source quality, so 300 scraper sites add noise, not signal — and they frequently introduce NAP inconsistencies you then have to clean up.
Press release distribution. Syndicated releases republished across dozens of low-quality sites are recognizable and discounted.
Fake community presence. Detected, and the downside is real.
Chasing Wikipedia. Almost no local service business meets notability requirements. Attempting it wastes weeks.
The realistic timeline
- Weeks 1–4: directories, associations, manufacturers, license verification. Most of the achievable gain is here.
- Months 2–3: press, roundups, sponsorships.
- Months 3–12: community presence, review accumulation, the compounding of consistent signals.
Entity building is the slowest AI visibility lever and the most durable. Schema can be replicated by a competitor in an afternoon. Fourteen independent sources describing you consistently over three years cannot.
What to do this week
- Search your business name in quotes and count independent mentions.
- Ask an assistant to describe your business and note every error.
- Claim and correct the nine major directories.
- Confirm your state license listing matches your public name.
- Claim every trade association and manufacturer listing you qualify for.
- Add every profile URL to your schema's
sameAs.
Then measure it
Entity strength shows up as being named at all, and as being described accurately when someone asks about you directly. Our free scan includes both — generic questions that test whether you exist in the market, and branded questions that test what the assistants know.
About fifteen seconds, free, no account.
Common questions
- What is an entity in practical terms?
- A distinct thing an AI system has a coherent picture of — your business, with a name, a location, services, credentials, and a reputation. The picture is assembled from every source that mentions you. Consistency across sources makes it sharp; contradictions and absences make it blurry.
- How do I know if my entity is thin?
- Search your business name in quotes and count results that are not your own properties. Fewer than about eight independent mentions is thin. Then ask an assistant to tell you about your business — vagueness or errors in the answer show you exactly where the gaps are.
- Does a Wikipedia page help?
- It would, but almost no local service business meets Wikipedia's notability requirements, and attempting a page you do not qualify for wastes time and can backfire. Focus on the achievable sources: directories, associations, manufacturers, local press, and community mentions.
- How long does entity building take?
- Directory and association listings show up within weeks. Press mentions, community discussion, and the accumulation of consistent signals take months. It is the slowest of the AI visibility levers and the most durable once built.
See exactly what the five AIs say about your business.
Twenty-five real customer questions across Claude, ChatGPT, Gemini, Perplexity, and Google AI Overviews. Free, about fifteen seconds, no account.
Run the free scanKeep reading
- The 15-Point AI Visibility Audit You Can Run YourselfTechnical fixes
- FAQPage Schema: The Fastest Path Into AI AnswersTechnical fixes
- Getting Mentioned on Reddit (and Why AI Weighs It So Heavily)Technical fixes
- A 30-60-90 Day AEO Plan for a Local Service BusinessMeasurement & buying
- What Is Answer Engine Optimization? A Local Business Owner's GuideFoundations