Why AI Search Engines Struggle to Identify Brands
When a large language model like ChatGPT or Perplexity tries to answer a question about your brand, it does not simply look up your website. It draws on a web of signals gathered during training and from live retrieval: Wikipedia entries, social media profiles, press coverage, structured data, and third-party knowledge bases like Wikidata and Google's Knowledge Graph. If those signals point in slightly different directions, or if some of them are missing entirely, the AI gets confused about who you actually are.
This is more common than most business owners realise. Two companies might share a similar name. Your brand might appear under slightly different spellings across platforms. Your LinkedIn page might describe you one way while your Crunchbase listing says something else. To a human reader, these inconsistencies are easy to gloss over. To a language model trying to build a reliable picture of your business, they are genuine problems.
That is where the sameAs property comes in. It is one of the most practical and underused tools in schema markup, and it directly addresses the question of identity disambiguation for AI systems.
What the SameAs Property Actually Does
The sameAs property is part of Schema.org's vocabulary, and it was designed to express a simple idea: this thing (your business, person, or product) is the same entity as the one described at these other URLs.
When you include a sameAs array in your Organisation or Person schema, you are telling any machine reading your page: "Do not guess. Here is the authoritative list of places where you can find verified information about us." For an AI model doing entity resolution, that list of URLs acts like a map. It connects your website to your Wikidata entry, your LinkedIn profile, your Companies House listing, your Twitter/X account, and anywhere else you maintain a confirmed presence.
The result is a cleaner, more confident entity match. When an AI summarises your brand or recommends you in a response, it draws on a unified set of signals rather than hedging because it cannot be sure which "Acme Solutions" it is actually talking about.
Which URLs Should You Include in SameAs
Not every URL is worth adding. The sameAs property works best when it points to authoritative, persistent, machine-readable sources. Here is a practical breakdown of what to prioritise.
Knowledge Graph and Encyclopaedia Sources
These carry the most weight because language models are trained heavily on structured data from knowledge bases.
- Wikidata: If your brand has a Wikidata entry (e.g.
https://www.wikidata.org/wiki/Q12345678), include it. This is arguably the single most valuablesameAsURL you can add. Wikidata is used directly by many AI systems as a reference for entity data. - Wikipedia: If a Wikipedia article exists for your brand, include it. Not every business will have one, but if yours does, use it.
- Google Knowledge Graph: You can include your Google Knowledge Panel URL if your entity has one, though Wikidata is the stronger signal.
Social Media Profiles
Social profiles are widely indexed and frequently appear in training data for LLMs. Include the ones where your brand is genuinely active and consistently named.
- LinkedIn company page
- Twitter/X profile
- Facebook business page
- Instagram profile
- YouTube channel
A word of caution: only include profiles that are active and accurate. A dormant Facebook page with outdated information can introduce inconsistencies, which is the opposite of what you want.
Business Directories and Registries
- Companies House (for UK businesses):
https://find-and-update.company-information.service.gov.uk/company/YOURCOMPANYNUMBER - Crunchbase company profile
- Companies registered with the relevant trade bodies in your industry
Press and Media Profiles
If your brand has a consistent profile on a major media platform, such as a Bloomberg company page or a CrunchBase listing, these are worth including. They reinforce the authoritative picture of your entity.
How to Write SameAs Schema in JSON-LD
The implementation is straightforward. Here is a basic example for an Organisation:
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "FlinnSchema",
"url": "https://flinnschema.com",
"logo": "https://flinnschema.com/logo.png",
"sameAs": [
"https://www.wikidata.org/wiki/Q12345678",
"https://www.linkedin.com/company/flinnschema",
"https://twitter.com/flinnschema",
"https://www.facebook.com/flinnschema",
"https://www.crunchbase.com/organization/flinnschema"
]
}A few important details:
- The
sameAsproperty accepts either a single URL string or an array of strings. Always use an array when you have more than one URL, which you almost certainly will. - Each URL should point to a page that is specifically about your brand, not your brand's homepage or a generic directory listing page.
- Use canonical, permalink-style URLs. Avoid URLs that might change, redirect, or expire.
For a Person entity, the structure is identical. If you are a consultant, author, or public-facing expert, you can attach sameAs links to your personal LinkedIn, your Google Scholar profile, your ORCID if relevant, and any Wikipedia or Wikidata entries about you.
If you are unsure where your JSON-LD block should sit in the page structure, the post on where JSON-LD schema should go in your HTML covers that in detail.
SameAs vs Other Identity Signals: How They Work Together
The sameAs property does not operate in isolation. It works best when it reinforces a consistent set of signals across your website and external platforms. Think of it as the connective tissue between all the other things that help AI identify and trust your brand.
Organisation Schema as the Foundation
Your sameAs array should always sit inside a well-formed Organisation or Person schema block. That parent schema establishes who you are (your name, URL, description, logo, contact details). The sameAs property then extends that identity outward to the rest of the web. Without the parent schema, the sameAs links have no anchor point.
The post on the difference between Organisation and LocalBusiness schema is worth reading if you are unsure which entity type applies to your business.
Consistency Across Profiles
Your schema can declare that your LinkedIn profile is the same entity as your website, but if your LinkedIn bio uses a different trading name, describes different services, or lists a different location, that inconsistency undermines the signal. Before you implement sameAs, audit the profiles you plan to include. Make sure the name, description, and core details are consistent across all of them.
This is one of the things the team at FlinnSchema checks during an AI Visibility Audit: not just whether the schema is technically correct, but whether the signals it points to are actually coherent and consistent.
Author Schema for Individual Experts
If your content marketing strategy relies on individual authors who are positioning themselves as experts, sameAs on a Person schema is equally important. An author with a Wikidata entry, a LinkedIn profile, and a ORCID identifier linked via sameAs sends a much stronger trust signal to AI systems than an anonymous byline. You can read more about that approach in the guide on using Author schema to build AI trust in your content.
Common Mistakes That Undermine SameAs
A few patterns come up repeatedly when reviewing schema implementations that are not performing as expected.
Including Your Own Homepage
Some implementations include the brand's own website URL in the sameAs array. This is unnecessary and slightly confusing. The url property already declares your website. The sameAs property is specifically for external third-party sources that confirm your identity from the outside. Keep them separate.
Using Profile URLs That Redirect or Break
Social media platforms occasionally change their URL structures. A LinkedIn company URL that worked perfectly in 2022 might redirect to a slightly different path today. Check your sameAs URLs periodically and update them if they have changed. A broken link in your schema does not cause an error, but it loses you the signal that URL was meant to carry.
Listing Profiles With Inconsistent Brand Information
As mentioned above, an inconsistent profile is worse than no profile. If your Facebook page still shows your old business name from before a rebrand, either update the page or remove it from your sameAs list until it is corrected.
Forgetting to Validate
After making any change to your schema, run it through Google's Rich Results Test and Schema.org's validator to confirm there are no syntax errors. A malformed JSON-LD block can cause the entire schema to be ignored. The guide on how to validate your schema markup before publishing walks through the exact steps.
The Bigger Picture: Entity Authority in AI Search
The underlying concept here is entity authority. AI search engines do not just rank pages; they recognise and evaluate entities. A brand that has a clear, consistent, well-documented entity presence across the web is one that an AI system can confidently cite, recommend, and describe.
The sameAs property is one of the most direct ways to build that entity authority. It takes maybe 20 minutes to implement properly, and the upside is that every AI system that reads your page gets a clear, unambiguous picture of exactly who you are and where to find verified information about you.
For brands that want to be cited in ChatGPT responses, Perplexity summaries, or Gemini answers, this kind of foundational identity work is not optional. It is the baseline from which everything else follows.
Frequently Asked Questions
Does SameAs schema directly affect Google rankings?
Not directly in the traditional sense of ranking signals. Google uses sameAs primarily for entity disambiguation and Knowledge Graph enrichment rather than as a direct ranking factor. That said, a well-defined entity with strong sameAs connections is more likely to earn a Knowledge Panel, which has real visibility benefits.
How many URLs can I include in a SameAs array?
There is no hard limit in the Schema.org specification. In practice, somewhere between four and ten high-quality, consistent URLs is a sensible target. Quality matters far more than quantity. Ten strong, consistent profiles will outperform twenty inconsistent or low-authority ones.
Do I need a Wikidata entry to use SameAs effectively?
No. Wikidata is the most powerful individual source you can include, but plenty of brands build strong entity signals without it. If your business does not yet have a Wikidata entry, focus on social profiles, business registries, and industry directories. Creating a Wikidata entry is worth doing if your brand meets Wikidata's notability criteria, but it is not a prerequisite for the rest of your schema work.
Should I include SameAs on every page or just the homepage?
Your core Organisation schema (including sameAs) should appear on every page, or at minimum on the homepage and key landing pages. Many brands include it in a sitewide template so it is automatically present everywhere. This is the right approach. There is no penalty for including it on multiple pages, and the consistent presence reinforces the signal.