sameAs, KnowledgePanel, and Entity Linking: Connecting Your Brand Across the Web for AI

Search engines and AI platforms no longer match keywords to pages. They match entities to answers. When someone asks ChatGPT for a recommendation, queries Perplexity about your industry, or triggers a Google AI Overview, the system behind that response isn't scanning for your URL. It's checking whether it can confidently identify who you are, verify that identity against multiple sources, and trust you enough to cite.

In mid-October 2025, ChatGPT rolled out a major update that changed how brands appear in its answers. According to Profound's analysis of millions of prompts, brands are now mentioned less frequently, and the competition for visibility has tightened.

ChatGPT began consistently tagging brands with a structured entity field, and the average number of brand mentions per answer dropped from about 6–7 to 3–4. The slot machine got smaller. If AI systems can't resolve your brand as a distinct, trustworthy entity, you don't just lose one ranking position - you lose the citation entirely. This guide walks through the practical infrastructure of entity identity: the sameAs property in Schema.org markup, how it connects to Knowledge Panels and knowledge graphs, and why entity linking has become the connective tissue between your brand and every AI platform that might recommend you.

What Entities Actually Are (And Why AI Systems Need Them)

An entity is not a page. It's not a keyword. To Google, an entity is a distinct, identifiable person, place, organization, event, or concept with specific attributes, clearly defined relationships, and a recognizable identity within its Knowledge Graph.

That last phrase - "recognizable identity" - is the one practitioners underestimate. We're witnessing a three-stage evolution in how the web is indexed and understood. Phase 1 (Strings): Traditional SEO optimized for keyword strings. Phase 2 (Things): Modern search understands entities. Knowledge graphs allow engines to recognize that a brand, a founder, and a product are distinct, related "things." We're now deep into Phase 2, with Phase 3 - where AI agents autonomously act on entity data - emerging fast. The practical implication is direct. AI systems cross-reference signals from multiple sources and formats. Your brand description on LinkedIn should align with what appears on your site. Profiles on Crunchbase, review platforms, or industry directories should reinforce the same category, positioning, and value proposition. When these signals are consistent across sources, AI systems can categorize and reference your brand with greater confidence. When they conflict, confidence drops, and your brand is less likely to be mentioned.

Entity identity isn't a marketing concept. It's an engineering problem. And sameAs is the first tool you use to solve it.

The sameAs Property: Your Brand's Machine-Readable ID Card

The sameAs property in Schema.org does exactly what its name suggests: it tells machines that two or more URLs refer to the same real-world entity. It allows search engines to determine that two or more web pages or social media accounts are referring to the same entity. This schema markup is typically used for businesses, public figures, and organisations.

When you add a sameAs array to your Organization or Person schema, you're creating a declarative identity layer. The sameAs property was created to state that an entity is exactly the same as an entity from another source. Using the sameAs property means the entity you're marking up inherits all of the same information, attributes and relationships of the external source. It is often used when an Organization, Brand or Person also exists on Wikipedia or Wikidata, or has social media sites.

Here's what a properly structured sameAs implementation looks like in JSON-LD:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://example.com/#organization",
  "name": "Example Company",
  "url": "https://example.com",
  "sameAs": [
    "https://www.linkedin.com/company/example",
    "https://www.wikidata.org/wiki/Q12345678",
    "https://en.wikipedia.org/wiki/Example_Company",
    "https://www.crunchbase.com/organization/example",
    "https://twitter.com/example"
  ]
}

What URLs Belong in sameAs (and What Doesn't)

Not every URL deserves inclusion. The sameAs property should only be used to relate entities that are truly equivalent, such as different versions of the same webpage or the same person's profile on different social media platforms. Listing competitor URLs, vaguely related industry pages, or unrelated entities actually harms disambiguation rather than helping it.

You should ensure that the URLs included in your sameAs property are valid and accessible since a URL returning a 404 error could negatively impact your SEO. Dead links in your sameAs don't just fail silently - they introduce noise into the identity resolution process. Priority URLs for most organizations:

• Wikidata entity page (highest impact for AI systems)
• Wikipedia article (if one exists)
• LinkedIn company page
• Crunchbase profile
• Official social media profiles (X/Twitter, Facebook, YouTube)
• Industry-specific directories (G2, Clutch, professional registries)

One common mistake is using Google Knowledge Graph Machine IDs (KGMIDs) as sameAs values. Google's John Mueller said technically you can use the knowledge graph ID URL for the URL in your sameAs property but it is not recommended. It is not recommended because that ID might change, so you probably want to use URLs that are less likely to change. Use stable, external URLs - Wikipedia, Wikidata, your company LinkedIn - instead of Google-generated identifiers.

Knowledge Panels: The Visible Proof That Google Recognizes You

A Knowledge Panel is the visible output of Google's entity understanding. Knowledge panels are information boxes that appear on Google when you search for entities (people, places, organizations, things) that are in the Knowledge Graph. They are meant to help you get a quick snapshot of information on a topic based on Google's understanding of available content on the web. Knowledge panels are automatically generated, and information that appears in a knowledge panel comes from various sources across the web.

Having a Knowledge Panel doesn't just improve your brand SERP. It creates a compounding feedback loop for AI visibility. When a Wikidata item feeds into Google's Knowledge Panel, that panel appearance creates additional training signal for LLMs that scrape search results. The Panel becomes a trust node that other systems reference.

How to Earn One

Panels can't be purchased or forced into existence. You earn it by proving you're a real entity with consistent, verifiable facts across the web. The process requires what practitioners call a "same facts trail" - every important attribute of your brand backed by at least two independent, trusted sources. Concrete steps that move the needle: 1. Establish an Entity Home - a single authoritative page (usually your About page or homepage) that defines your brand clearly. It is the specific URL that Google (and other AI platforms) reconcile all other mentions back to. When Google sees your name mentioned on 500 different websites, it needs a single reference point to confirm which entity all those mentions refer to. The Entity Home is that reference point.

2. Create a Wikidata entry - there is no notability requirement, no editorial review. It is accessible to any brand today. Populate it with your official name, industry classification, founding date, primary URL, and geographic location. This alone shifts the probability of entity recognition. 3. Standardize NAP and brand descriptions across every platform. Standardize NAP for local entities (name, address, phone) across your website, GBP, and major directories. Use the same headshot and logo files across key profiles, so Google sees stable image signals.

3. Earn independent editorial coverage - interviews, industry publications, association memberships - that corroborates your entity facts from third-party sources.

Claiming Your Panel

Once a Knowledge Panel appears, claiming it gives you limited but valuable control. Claiming your knowledge panel gives you a degree of control over its content. Google accepts feedback from the person who claimed a knowledge panel. You can suggest edits, update images, and - critically - merge different Google entities in the Knowledge Graph. A successful merge will result in more source references pointing to a single entity in the graph. The number of sources supporting an entity in Google's knowledge graph is a trust signal.

Entity Linking: The Bridge Between Your Content and the World's Knowledge

Entity linking goes beyond sameAs on your Organization schema. In the context of Schema Markup, entity linking is the act of linking the entities on your site to the corresponding known entities on external authoritative knowledge bases such as Wikipedia, Wikidata and Google's Knowledge Graph using Schema.org properties. Examples of connector properties include sameAs, mentions, areaServed, and more.

Where sameAs says "this organization is that entity," entity linking says "this page talks about these known concepts." It uses properties like about, mentions, and spatialCoverage to connect the topics within your content to their canonical definitions in external knowledge bases.

Real-World Results

The impact of entity linking is measurable, not theoretical. Schema App added spatialCoverage and audience properties to identify the state a page pertained to, then clarified which entity was being referred to by nesting equivalent entities defined on Wikipedia, Wikidata and Google's Knowledge Graph using the sameAs property. After running the experiment for 85 days, the test sites saw a 46% increase in impressions and a 42% increase in clicks for non-branded queries.

In a separate case, Schema App began implementing Entity Linking across a prioritized set of entities tied to their topic authority goal. When comparing AI Overview visibility, they observed a 19.72% increase in AI Overview visibility for entity-related keywords.

Marshfield Clinic Health System experienced an 80% increase in traffic by scaling Schema Markup across 50+ sites. They also saw a 454% increase in CTR when Review Snippet markup was awarded and a 32% increase in CTR on physician pages after implementing entity linking.

Internal Versus External Entity Linking

For Internal Entity Linking, you define, manage, and connect the unique entities that matter most to your organization-such as your products, services, people, and locations. By linking these internal entities across your website content, you can unify your brand-specific knowledge and build a consistent Content Knowledge Graph. Together, internal and external entity linking enable enterprises to strengthen their knowledge graphs, improve content discoverability, and future-proof their SEO strategies.

The @id property is what makes internal linking work. The @id property creates unique references within your structured data, while sameAs links connect to authoritative external profiles. This combination helps search engines confidently identify who's who, even when dealing with common names or multiple authors.

A common implementation error: re-declaring your full Organization entity on every page instead of referencing the homepage entity via @id. Every page re-declares the full LocalBusiness entity instead of referencing the homepage entity via @id. This creates multiple conflicting entities instead of one authoritative one.

Why This Matters More for GEO Than Traditional SEO

Generative Engine Optimization exists because the discovery layer has fragmented. GEO is the practice of positioning your brand and content so that AI platforms like Google AI Overviews, ChatGPT, and Perplexity cite, recommend, or mention you when users search for answers. The stakes are real: ChatGPT reaches over 800 million weekly users. Google's Gemini app has surpassed 750 million monthly users. And AI Overviews are appearing in at least 16% of all searches. More recent data suggests that percentage has climbed substantially. AI Overviews now appear in 25.11% of Google searches, up from 13.14% in March 2025.

Entity linking is the mechanism that makes your brand citable across all of these surfaces simultaneously. Here's why: AI systems perform multi-source verification. GEO engines rely on language models that evaluate entity coverage, factual consistency, and cross-source agreement before deciding whether to cite you. When your structured data connects to Wikidata, your LinkedIn matches your website, and your Crunchbase profile reinforces the same positioning, every verification check resolves cleanly. Structured data now acts as a trust signal, not just a display trigger. Google's Gemini-powered AI Mode uses schema markup to verify claims, establish entity relationships, and assess source credibility during answer synthesis. Schema that accurately describes content increases the probability of AI Mode citation even when no traditional rich result is displayed. This is the most important strategic shift in structured data since rich snippets launched.

Citation slots are shrinking. The average number of brand mentions per answer dropped from about 6–7 to 3–4. Average brand visibility fell by approximately 31% across tracked brands, with more than 85% seeing declines. Fewer slots means every disambiguation signal compounds in value.

Wikidata: The Free Lever Most Brands Ignore

If there is a single underinvested asset in the GEO stack, it's Wikidata. Wikipedia is a human-readable encyclopedia that feeds into LLM training data and builds topical authority in AI models. Wikidata is a machine-readable structured database queried directly by AI systems and search engines to resolve entity facts.

Every major AI system-ChatGPT, Gemini, Claude, Apple Intelligence-uses Wikidata for factual grounding. Your Wikidata entry, or the absence of one, directly influences how AI systems understand your brand.

The barrier to entry is remarkably low. Wikidata is the higher-priority starting point because it has no notability requirement and is immediately machine-readable. Yet most marketing teams don't know it exists.

Setting Up a Wikidata Entry That Actually Works

Creating the entry is the easy part. Creating a Wikidata item takes an afternoon. Maintaining it takes years. This asymmetry is where most visibility strategies fail.

The five priority properties for any Wikidata item are instance of (P31), official website (P856), inception date (P571), external identifiers, and sitelinks to Wikipedia or Wikimedia Commons. Beyond these basics, add identifiers from business registries. Organizational entries gain trust by triangulating with external registries such as OpenCorporates and GLEIF, building what functions as a multi-source trust stack.

Maintenance matters because duplicate items fragment facts, causing LLMs to hallucinate by mixing information from both entries. Missing references cause statements to lose trust weight. Outdated identifiers create dead ends that retrieval systems interpret as signal decay. Inconsistent labels and aliases confuse disambiguation at the exact moment the model is deciding which entity to cite.

Once your Wikidata entry exists with a Q number, that URL becomes the most powerful value in your sameAs array.

The Implementation Playbook: From Audit to Ongoing Governance

Entity linking and sameAs implementation isn't a one-time project. At enterprise scale, manual updates are a liability. You must treat schema as an ongoing operational discipline.

Step 1: Audit Your Entity Footprint

Search your brand name across Google, ChatGPT, Perplexity, and Claude. Document what each system says. Look for inconsistencies, incorrect attributions, and missing connections. Brightview Senior Living's brand was frequently confused with unrelated entities in AI-generated and local search results, undermining trust and discoverability. Through Schema App's Entity Linking feature, Brightview was able to accurately associate their brand and locations with the correct entities in Google's Knowledge Graph. You can't fix what you haven't mapped.

Step 2: Build Your Entity Home

Designate one URL as your canonical entity reference. Place comprehensive Organization schema on it, including @id, name, legalName, alternateName, foundingDate, url, logo, and a full sameAs array. Every other page on your site should reference this entity via @id, not re-declare it.

Step 3: Create or Audit Your Wikidata Entry

Check whether your organization has a Wikidata item. If not, create one. If it exists, verify every statement has a reference. Look for duplicates. Wikidata is the single highest-impact external platform for entity authority, and it is free to claim.

Step 4: Implement Entity Linking Across Content Pages

For every core topic you publish about, identify the canonical entity in Wikidata or Wikipedia. Use about and mentions properties in your Article or WebPage schema to link to those entities. Establish one primary entity per page and identify 3–6 supporting entities that provide relevant context. Connect these entities to Wikipedia, Wikidata, industry standards, and your pillar content through strategic internal and external linking.

Step 5: Validate and Monitor

Run your pages through the Schema.org validator, not just Google's Rich Results Test. The Schema.org validator validates more than what's available as a rich result in Google Search. Set up quarterly audits to check for broken sameAs URLs, inconsistent entity declarations, and schema drift - the gap between what your visible content says and what your structured data claims.

At enterprise scale, the greatest threat to visibility is schema drift. This occurs when your human-visible content evolves, but your machine-readable schema remains static. When AI systems detect this inconsistency, they lower your confidence score. Reduced confidence leads to zero citations.

What Happens When You Get This Right

The brands that treat entity identity as infrastructure - not as a one-time SEO tactic - compound their advantage over time. The brands capturing those 3-4 citation slots today are building advantages that compound over time - more visibility leads to more mentions, which strengthens entity recognition, which increases future visibility.

An AccuraCast study analyzing over 2,000 prompts across ChatGPT, Google AI Overviews, and Perplexity found that 81% of web pages receiving citations included schema markup. That's not a correlation you can afford to ignore. Source citation improves by 30% when schema markup is included.

The specifics of how each AI platform weights structured data will continue to evolve. What won't change is the underlying principle: machines need to resolve your identity before they can trust your content. The sameAs property, a robust Knowledge Panel, and systematic entity linking are how you make that resolution clean, consistent, and impossible to confuse with a competitor. Start with Wikidata. Build your Entity Home. Connect your sameAs. Then keep the data fresh - because in a world where AI citations decay within weeks, the brands that maintain their entity infrastructure are the ones that keep showing up.