One of the biggest challenges in AI search right now is helping large language models understand who you are and what you do. If AI models are getting your brand wrong, advanced schema markup can fix that. In this post, we’re covering three schema properties that help disambiguate your brand and improve how you show up across search platforms.
AI models are already making decisions about your brand. The question is whether they have enough information to get it right. You can use structured data to help AI models understand your brand.
How Well Do AI Models Know Your Brand?
AI models don’t parse your intent the way you might think. They resolve entities probabilistically. They take inputs and predict the next best response based on training data. When that data is incomplete, AI models query search engines and other databases to fill the gaps.
Why Does AI Keep Getting My Brand Wrong?
The problem isn’t always content quality; it’s content clarity. If the signals around your brand are unclear, the model will make its best guess and might be wrong.
What Causes Entity Confusion
A few common culprits:
- Shared names. Your brand name might be similar to a competitor or a completely unrelated company in another industry.
- Category overlap. If your product touches multiple categories, that can create confusion about what you do.
- Ambiguous acronyms. Your brand abbreviation can be confused with another similar acronym.
A real-world example: a client in the cybersecurity networking space sells a product called SASE, which stands for Secure Access Service Edge. But SASE also refers to self-addressed stamped envelopes, and to The Society for the Advancement of Socio-Economics. Without clear signals, an AI model could pull any of those meanings when your content comes up.
3 Schema Properties That Resolve Ambiguity
The three schema properties that you’ll want to implement are sameAs, about, and mentions.
1. sameAs — Anchor to Authoritative Sources
SameAs connects your entity to external authoritative sources so models can understand who you are across the web. Think Wikipedia, LinkedIn, Crunchbase, G2, Gartner Peer Insights, Forrester, and your social profiles.
Wikipedia and Wikidata carry a lot of weight, especially in ChatGPT. If you have a Wikidata entry, link to it. If you don’t, getting one approved today is a significant lift and probably not worth the effort for most brands. Instead, build your sameAs corpus from the profile pages you already control, and add those links within your organizational schema. Those sources often carry more authority than your own domain, with some models, and building those links into your schema is what creates the bridge.
2. about — Declare Topical Identity
The about property declares the primary topical identity of your site or a specific page. It works at both the website and page levels.
This is where Wikidata QIDs come in. A QID is a linked open data point that creates a machine-readable relationship between the keyword you’re declaring and the entity itself. Instead of just using words to describe your topic, you’re using structured data to create an entity bridge that search engines and AI crawlers can actually process.
One important distinction: This is different from the Wikidata brand page mentioned in sameAs. You don’t need a Wikidata page for your company to use QIDs here. These are existing entries for topics and concepts — such as “Secure Access Service Edge” or “cloud security” — that already live in Wikidata. You’re simply finding the QID for the topic your content covers and referencing it. No approval process, no heavy lift.
3. mentions — Build the Topical Neighborhood
Scheme markup for mentions extends the about property by identifying secondary entities referenced in your content. If about is your topic, mentions are the related terms that surround it.
Going back to the SASE example: an article might declare it’s about as Secure Access Service Edge, then use mentions to include SD-WAN, ZTNA, and cloud security. This builds your topical neighborhood, placing your content on a conceptual map that the model can navigate. Linked data is a powerful lever here. Not just links, but data that tells the machines exactly what you mean.
Schema Examples (Live JSON-LD Demo)
Organizational Schema with sameAs
The first JSON-LD block covers organizational schema with sameAs links added. This includes the company name, URL, description, and a set of sameAs links pointing to the company’s Wikidata entry, LinkedIn, Crunchbase, G2, and product pages. Each of those links creates a connection back to your domain and helps AI crawlers map the relationships around your brand.
Website and about Schema
The second block lives alongside the first in the same <head> tag. It declares the website type and uses the about property to define what the site is about, connected to a Wikidata QID. This isn’t just a hyperlink. It’s linked data that tells the machine that when we say “SASE,” we mean “Secure Access Service Edge.”
Article and mentions
The third block applies to blog posts and other content pages. The author points back to the organization or to a specific author ID if that person has credibility in the space. Each article declares its about topic, then lists mentions for all the related entities covered. You can also include alternative names to go even deeper. Check schema.org for the full range of options.
How To Get Started With Schema Markup
Most brands are running boilerplate schema. Someone installed a plugin, set it, and forgot about it. That is not how this works anymore. Here’s how to build it out properly:
- Start on your homepage. Get your website and organizational schema in place first.
- Obtain your QIDs. If you have a Wikidata entry, use it. If not, identify the terms you’re known for and pull the QIDs for each.
- Test your pages. Use schema.org and the Rich Results tool to make sure nothing is broken. Broken schema can be worse than no schema at all.
- Build into your product pages. Once your homepage is solid, extend the approach to your key product and service pages.
- Apply to blog content. Roll out article schema with about and mentions across your content.
- Keep validating. Test changes as you make them and monitor how AI search results respond, along with traditional search.
Entity equity compounds. The clarity you build now pays off over time. Start now, be as clear as possible, and don’t set it and forget it.
Schema markup has always mattered for search, but for AI search, it’s doing a different kind of work. It’s not just about rich results anymore. It’s about making sure AI models can accurately identify, understand, and surface your brand. The three properties covered here, sameAs, about, and mentions, give you a structured way to close the gap between what you know about your brand and what the models actually understand.
If you want help building a schema strategy for your site, contact us. We’d love to partner with you.
Until next time, happy marketing.