Entity SEO is the work of making sure search engines and AI assistants can tell exactly which business you are and attach the right facts to you: name, category, location, founders, products and official profiles. You can’t submit your brand to Google’s Knowledge Graph, and Wikidata is not a listing service. What you can do is state one consistent identity on your own site, link it to the profiles you control, keep your listings accurate, claim the panels and profiles that already exist, and earn independent coverage that describes you the same way.
Only some of this is documented as feeding AI answers. Google says AI Overviews work “in tandem” with its ranking systems and the Knowledge Graph, and Microsoft’s Bing guidelines say “clear entity definition improves grounding visibility and citation accuracy.” OpenAI, Perplexity and Anthropic don’t document using knowledge graphs at all; their assistants search the web, so the pages that describe you are your entity data there. For the wider discipline, start with What Is Generative Engine Optimization?
What an entity is, and what a knowledge graph does
An entity is a uniquely identifiable thing: a company, person, product, place or concept. A knowledge graph stores entities with an ID, attributes (founding date, headquarters, website) and typed relationships to other entities (founded by, located in, owned by).
Google introduced its Knowledge Graph in May 2012 as a move to “things, not strings.” Its help center describes it as “our database of billions of facts about people, places, and things,” drawn from public sources and licensed data (How Google’s Knowledge Graph works).
An entity can carry several identifiers: a Google Knowledge Graph ID such as /m/0dl567 (returned by Google’s Knowledge Graph Search API), a Wikidata QID if it has a Wikidata item, a schema.org @id you define in your own markup, and sameAs links in that markup pointing to your LinkedIn, Crunchbase, X or Wikidata pages. They all do one job: let a system be sure that “Acme” on your site, “Acme Inc.” on LinkedIn and “Acme” in a news story are the same company.
What each AI platform documents about entity data
Checked November 11, 2026, only Google and Microsoft say anything specific.
| Platform | What it documents | Source |
|---|---|---|
| Google AI Overviews and AI Mode | AI Overviews use a custom Gemini model that “works in tandem with our existing Search systems – like our quality and ranking systems and the Google Knowledge Graph.” AI Mode “taps into real-time information from the web and from Google, like facts from our Knowledge Graph.” | Google, AI Overviews and AI Mode in Search (May 2025) |
| Google (local and commerce) | AI responses “can include product listings, product information, and information about local businesses”; Merchant Center feeds and Google Business Profiles help you be visible in them | Google’s AI optimization guide, updated July 10, 2026 |
| Microsoft Bing and Copilot | ”Use clear and consistent naming for people, organizations, products, and locations… Clear entity definition improves grounding visibility and citation accuracy.” | Bing Webmaster Guidelines |
| OpenAI (ChatGPT) | Search “sometimes partners with other search providers” and may use third-party providers for local information such as business listings. Nothing on knowledge graphs or entity markup. | Searching the web with ChatGPT |
| Perplexity | Crawler documentation only | Perplexity crawlers |
| Anthropic (Claude) | Crawler documentation only (ClaudeBot for training, Claude-SearchBot for search, Claude-User for user fetches) | Anthropic help center |
Even at Google the Knowledge Graph works alongside retrieval, not instead of it: the July 2026 guide says its AI features rely on “our core Search ranking systems to retrieve relevant, up-to-date web pages from our Search index.” For ChatGPT, Perplexity and Claude, entity data means what the pages they retrieve say about you, plus what their models learned in training; if those pages disagree about your name, location or category, so may the answer. Engine-by-engine detail: Google AI Overviews and AI Mode, ChatGPT, Perplexity, Claude.
What the studies show
No controlled study has shown that a Wikidata item or a knowledge panel causes more AI citations. Three large observational studies are worth knowing.
Where AI answers get their facts about businesses. Yext analyzed 6.8 million citations from 1.6 million queries per model across ChatGPT, Gemini and Perplexity, collected July 1 to August 31, 2025, for retail, financial services, healthcare and food service locations (Yext Research, October 9, 2025). Brands’ own websites made up 44% of citations and listings 42%, with reviews and social another 8%. Gemini leaned on websites (52.1%), OpenAI on listings (48.7%). These were location-specific consumer queries, and Yext sells listings management, so read it as “for local queries, your site and listings are most of what gets cited.”
Mentions and AI visibility move together. Ahrefs compared 10,000 US brands’ mentions on sites they don’t control with how often they were named in answers to 260 million non-branded prompts on ChatGPT, AI Mode and AI Overviews (Louise Linehan and Xibeijia Guan, Ahrefs, November 3, 2026). Every mention type correlated strongly (Spearman): Reddit threads that rank in Google 0.75, webpage mentions 0.70, news articles 0.67, YouTube 0.64, TikTok 0.60. ChatGPT was the least tied to off-site mentions, and brand size explains part of the pattern: among brands with similar search demand, the correlation fell to 0.34–0.54. Ahrefs says it can’t show that mentions cause visibility.
Being named and being cited are different. Semrush’s October 2026 AI Visibility Index (fixed monthly sets of 25 million US and 2.5 million UK prompts on ChatGPT, Gemini and Google AI Mode, plus AI Overviews data) found that, averaged across 22 categories, fewer than half of the 100 most-mentioned brands were also among the 100 most-cited domains, in every month and both markets (Semrush, October 13, 2026). Being recognized as an entity and having your own site cited are separate goals; track both.
On schema, the best controlled test so far found no citation gain from adding JSON-LD; the evidence is in Structured Data and AI Citations.
Entity SEO vs. keyword SEO
| Keyword SEO | Entity SEO | |
|---|---|---|
| Unit of work | A page and its target queries | The business, its people and products |
| Where the work happens | Your pages and links to them | Your site plus profiles, listings, knowledge bases and coverage elsewhere |
| What goes wrong without it | You don’t rank | Engines confuse you with someone else, repeat stale facts, or don’t name you |
| How to measure it | Rankings, clicks | Appearance rate, citation rate, share of AI voice, and whether answers state your facts correctly |
Audit your entity footprint
Do this before changing anything, and keep the results as your baseline.
- Search your brand on Google. Is there a knowledge panel or Business Profile? Are the name, logo, website and description right? Is it mixed up with a namesake?
- Check the Knowledge Graph Search API. It returns matching entities with IDs and short descriptions. Google calls it read-only and “not suitable for use as a production-critical service,” and is moving it to Cloud Enterprise Knowledge Graph, so treat it as a spot check.
- Search Wikidata and Wikipedia for an item or article about you, or about a namesake, and check its facts.
- List every profile and listing: Google Business Profile, Bing Places, Apple Business Connect, LinkedIn, Crunchbase, Yelp, industry directories, review sites. Record the exact name, address, phone, URL and one-line description on each. That’s your inconsistency map.
- Ask the assistants about you. Run “What is [brand]?”, “Who founded [brand]?” and “Where is [brand] based?” on ChatGPT, Gemini, Perplexity, Claude and Google AI Mode. Note each wrong fact and the source cited; wrong facts usually trace to a specific page.
- Run your buyer prompts. A fixed set of category questions (“best [category] for [use case]”), each run more than once, recording whether you’re named, linked, or replaced by a competitor. How to Measure AI Citations covers sampling.
Make your own site the clearest source
State the facts in plain text. Your homepage and About page should say what the business is, who runs it, where it’s based, when it started, what it sells and who it serves, using one form of your name. Give founders and authors their own bio pages. Facts that exist only in markup may never reach an answer: in the fetch tests summarized in our structured data guide, no assistant showed it was reading schema as schema.
Add Organization markup once, on the homepage or About page. Google says it “can help Google better understand your organization’s administrative details and disambiguate your organization in search results,” that properties such as naics and iso6523Code are “used behind the scenes to disambiguate your organization from other organizations,” and that others can influence the logo in your knowledge panel. No properties are required (Organization structured data, updated September 8, 2026).
{
"@context": "https://schema.org",
"@type": "Organization",
"@id": "https://example.com/#organization",
"name": "Example Co",
"legalName": "Example Co, LLC",
"url": "https://example.com",
"logo": "https://example.com/logo.png",
"description": "Example Co makes scheduling software for veterinary clinics.",
"foundingDate": "2020-01-15",
"founder": {
"@type": "Person",
"name": "Jane Example",
"url": "https://example.com/about/jane-example"
},
"address": {
"@type": "PostalAddress",
"addressLocality": "Austin",
"addressRegion": "TX",
"addressCountry": "US"
},
"knowsAbout": ["veterinary practice management", "appointment scheduling"],
"sameAs": [
"https://www.linkedin.com/company/example-co",
"https://www.crunchbase.com/organization/example-co",
"https://x.com/exampleco",
"https://www.wikidata.org/wiki/Q00000000"
]
}
- Use the most specific subtype that fits (for example
OnlineStoreor aLocalBusinesstype), as Google recommends. - List only
sameAsprofiles that exist, are clearly yours and state the same facts. Add a Wikidata URL (Q00000000 above is a placeholder) only if an item about you already exists; otherwise delete that line. There’s no published “right number” of links. - Keep every marked-up fact visible on the page; Bing says markup “must accurately reflect visible content.”
- Point your Article and Person markup at the same
@id(publisher,worksFor). - Validate with the Rich Results Test and the Schema Markup Validator.
Knowledge panels: what you can and can’t do
You can’t request one. Google says knowledge panels “are automatically generated” (About knowledge panels) when “there is enough information available on the open web” (How Google’s Knowledge Graph works). There’s no form and no timeline, and a panel that used to show can stop showing, which Google says “isn’t something we can or would influence.”
You can claim one that exists. Google’s Get verified on Google process has the subject or an official representative click “Claim this knowledge panel” and sign in to one of the official sites or profiles Google lists, such as YouTube or Search Console. “Not all knowledge panels are claimable.” Verified owners can suggest edits; anyone can use the panel’s Feedback link.
Local businesses use Business Profile. Google says businesses “have special Business Profile panels and should use Business Profile to create, claim, and make suggestions to those.” For a business with a location or service area, a verified, complete Google Business Profile is the entity record Google points you to, and it’s one of the sources Google names for AI responses. Multi-location brands need one per location; see Multi-Location AEO and Local Business AEO.
Wikidata and Wikipedia: when they help and when they backfire
Wikidata is an open, CC0-licensed knowledge base that many systems reuse, which is why marketers now target it and why its editors push back.
The bar. Wikidata’s notability policy accepts an item that has a valid sitelink to a Wikimedia page such as a Wikipedia article, or “refers to an instance of a clearly identifiable conceptual or material entity that can be described using serious and publicly available references,” or fills a structural need.
How editors apply it to companies. The Self-promotion essay (an essay, not policy, last edited November 2026) says creating an item about your own organization “is strongly discouraged” and “is likely to be deleted,” that a business registry listing “does not by itself make a company notable,” and that “Wikidata does not control Google’s Knowledge Panels, and they are not a valid reason to create an item.” Its rule of thumb is at least three independent, reputable sources with substantial coverage. In September 2026 Wikidata’s Reducing promotional editing page reported “a sharp increase in promotional editing” because “people are being directed towards Wikidata as a way to establish an Item online and make it more visible to search engines and generative AI systems.” Paid editing must be disclosed.
What to do.
- No item and little independent coverage: don’t create one or pay someone to. Earn the coverage first.
- An item exists: you may correct clear factual errors but shouldn’t remove sourced claims; request anything substantive on the item’s talk page with a source.
- Substantial independent coverage, no item: one may qualify. Core statements are instance of (P31), official website (P856), inception (P571), headquarters location (P159), and industry (P452) or field of work (P101), each referenced, plus identifiers such as LinkedIn company or organization ID (P4264) and Crunchbase organization ID (P2088). Disclose any connection and let the community decide.
Wikipedia is stricter. Its company notability guideline requires “significant coverage in multiple reliable secondary sources that are independent of the subject,” and doesn’t count press releases, sponsored posts, or “case studies” and “success stories” from consulting firms. Most small businesses don’t qualify yet.
Keep listings and profiles consistent
Listings were 42% of citations in Yext’s data and nearly half of OpenAI’s, so listing accuracy is some of the most direct entity work you can do.
- Use one exact name, address, phone number and URL across Google Business Profile, Bing Places, Apple Business Connect, Yelp, LinkedIn, Crunchbase and your industry’s main directories, in one format.
- Use the same one-sentence description and category everywhere. “Marketing agency” here and “software company” there is how assistants end up describing you wrongly.
- Fix or close duplicate and outdated listings, especially after a move, rebrand or merger, and keep hours, services and prices current.
- Treat reviews as entity data; see Review Management for AI Visibility.
Earn independent mentions, not inauthentic ones
Ahrefs found that widely discussed brands tend to be named more in AI answers, and that mentions on modest sites tracked visibility about as closely as press coverage. Google adds a warning in its July 2026 guide: its AI features “can show what’s being said about products and services across the web, including in blogs, videos, and forum discussions. However, seeking inauthentic ‘mentions’ across the web isn’t as helpful as it might seem,” because its ranking systems favor high-quality content and other systems block spam.
The useful work is what you’d do without AI: expert commentary for journalists and industry publications under a named author; podcast and event appearances; accurate information on the review sites and “best [category]” roundups in your space; honest reviews from real customers; taking part in buyer forums as yourself, with disclosure; and original data others want to cite. Scripted posts and fake testimonials are exactly the inauthentic mentions Google means. Note that a news story naming you without a link can still be retrieved and quoted, while a link from a page that never describes you does little for your entity.
A 30-day entity cleanup plan
This is a sequence, not a promise: Google publishes no timeline for knowledge panels, and no AI platform says how fast it picks up changes.
Week 1: audit. Run the audit above. Write down your canonical name, one-sentence description, category, address and founding facts. Save the assistant answers and prompt results as your baseline.
Week 2: your site. State those facts in text on the homepage and About page. Add or fix Organization markup with an @id and sameAs links to profiles that exist. Add Person pages for founders and authors. Validate.
Week 3: profiles and listings. Claim or verify Google Business Profile (if you have a location or service area), Bing Places and Apple Business Connect. Make every listing match your canonical facts and close duplicates. Claim your knowledge panel if one exists and is claimable; request corrections to an existing Wikidata item on its talk page.
Week 4: coverage plan and re-check. List the publications, podcasts, review sites and communities your buyers use, and plan how to earn genuine coverage there this quarter. Re-run the assistant questions and trace each remaining wrong fact to its source.
Then re-run the same prompt set monthly against the baseline. Expect slow movement; How Long Does AEO Take? covers what’s realistic.
How to measure entity visibility
Use the same definitions as How to Measure AI Citations:
- Appearance rate: the share of answers to your prompt set that name you.
- Citation rate: the share that link to your site.
- Share of AI voice: your share of brand mentions across the competitors you track.
- Fact accuracy: whether each assistant gets your name, category, location and founders right on branded questions, and which source it cites when it doesn’t.
Google Search Console’s generative AI performance report and Bing Webmaster Tools’ AI Performance report add impressions and citations on those platforms; tools are compared in Best AEO Tools in 2026.
Prefer to self-assess first? Download the 25-item AI Citation Readiness Scorecard (PDF), a printable self-assessment across schema, content architecture, authority, measurement and conversion. Free, instant download, no email required.
Want it done for you? The free AI Visibility Snapshot tests 10 prompts on ChatGPT, Google AI Overviews and Perplexity, with a short note on where you’re cited and where you’re not. The Citation Audit tests 30–50 locked prompts across six surfaces (ChatGPT, Google AI Overviews, Perplexity, Gemini, Claude and the OpenAI Responses API) and delivers competitor gaps, schema and content fixes and a prioritized 60–90 day roadmap in 5 business days. It’s $1,500, fully credited toward an Implementation Sprint or retainer signed within 30 days. We don’t guarantee citations or rankings. See pricing.
Request a free Snapshot or a Citation Audit →
FAQ
What is entity SEO?
Making sure search engines and AI assistants can identify your business as one distinct entity and attach the correct facts to it. It covers how you describe yourself on your site, your Organization markup and sameAs links, your listings, any knowledge panel or Wikidata item about you, and how independent sources describe you.
Do AI assistants use Google’s Knowledge Graph?
Google says its own do: AI Overviews work in tandem with its ranking systems and the Knowledge Graph, and AI Mode draws on Knowledge Graph facts. OpenAI, Perplexity and Anthropic don’t document using any knowledge graph; their assistants search the web and cite the pages they retrieve.
How do I get a Google knowledge panel?
You can’t request one. Google generates panels automatically when there’s enough information about an entity on the open web, with no published timeline. If a panel about you exists, you can claim it and suggest edits. Businesses with a location or service area should use Google Business Profile instead.
Should I create a Wikidata item for my company?
Usually not yourself. Wikidata requires a clearly identifiable entity described by serious, publicly available references, and its self-promotion essay says items created by the organizations they describe are likely to be deleted. If substantial independent coverage exists, an item may qualify; if one exists already, request corrections on its talk page.
Do brand mentions matter more than backlinks for AI answers?
They correlate more strongly. In Ahrefs’ 2025 study of 75,000 brands, branded web mentions correlated with AI Overview visibility at 0.664 versus 0.218 for backlinks. These are correlations; Ahrefs’ 2026 follow-up found brand size explains part of the link, and Google warns that seeking inauthentic mentions isn’t as helpful as it might seem.
Sources
- Google. AI Overviews and AI Mode in Search (PDF), May 2025.
- Google Search Central. Optimizing your website for generative AI features on Google Search. Last updated July 10, 2026.
- Google Search Central. Organization structured data. Last updated September 8, 2026.
- Google Knowledge Panel Help. How Google’s Knowledge Graph works, About knowledge panels and Get verified on Google.
- Google for Developers. Knowledge Graph Search API.
- Singhal, A. (May 16, 2012). Introducing the Knowledge Graph: things, not strings. Google.
- Microsoft Bing. Bing Webmaster Guidelines.
- OpenAI Help Center. Searching the web with ChatGPT.
- Perplexity. Perplexity crawlers.
- Anthropic Help Center. Does Anthropic crawl data from the web, and how can site owners block the crawler?
- Wikidata. Notability (policy), Self-promotion (essay) and Reducing promotional editing (September 2026).
- Wikipedia. Notability (organizations and companies).
- Yext Research (October 9, 2025). AI Citations, User Locations, & Query Context; source shares and model differences from Yext’s press release.
- Linehan, L. & Guan, X. (November 3, 2026). Which Types of Mentions Correlate the Strongest With AI Visibility? (260M AI Prompts Analyzed). Ahrefs.
- Linehan, L. (May 26, 2025). An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied). Ahrefs.
- Semrush (October 13, 2026). Semrush Releases 2026 AI Visibility Index, Revealing Its First UK Benchmark.
- Schema.org Organization, Google Rich Results Test, Schema Markup Validator.
All sources checked November 11, 2026.
Related reading: Structured Data and AI Citations, How to Measure AI Citations, The AI Citation Readiness Checklist, How to Optimize for Google AI Overviews and AI Mode, GEO Glossary.