The claim vs. the reality
Since AI search became a priority for agencies, structured data has been frequently cited as an important optimization lever. The logic sounds reasonable: if you mark up your content with schema.org vocabulary, AI models can "understand" it better and are more likely to surface the brand in their responses.
The reality is more nuanced. And more useful to understand properly.
How AI models actually consume web content
Large language models like GPT-4o are trained on text corpora that include the web. But structured data (JSON-LD, microdata) is primarily consumed by Google's crawlers and indexers - not by the training pipelines that shape what LLMs know about brands.
When OpenAI trains on Common Crawl or other web corpora, it's processing text. A JSON-LD block declaring that a business is a LocalBusiness with a priceRange property is visible in the HTML, but the primary signal that shapes ChatGPT's brand associations comes from the natural language text surrounding it - not the schema markup itself.
What structured data does help with
That's not to say schema markup is useless for AI visibility. There are indirect effects worth understanding:
Google AI Overviews - direct benefit
Direct benefitGoogle AI Overviews do use structured data as a ranking signal. FAQ schema, HowTo schema, and Article schema with clear authorship metadata are all relevant for AI Overview inclusion. This is the one channel where structured data has a documented direct benefit.
Traditional SEO - indirect benefit
Indirect benefitBetter structured data improves traditional Google rankings, which improves the overall authority and coverage of a brand on the web. More organic visibility → more third-party coverage → more mentions in the corpora AI models train on. The path is indirect but real.
Entity disambiguation - minor benefit
Minor benefitSchema markup that clearly identifies a brand as an Organization with a specific sameAs reference to Wikidata or other authority sources may help disambiguate the entity in AI training data. This is a real but small effect.
ChatGPT and Perplexity - minimal direct benefit
Minimal direct benefitFor ChatGPT specifically, schema markup on your own site has minimal direct impact on whether the model recommends your brand. GPT-4o's recommendations are shaped by training data breadth, not by parsing live website schema.
Where to spend the time instead
If you're deciding where to allocate optimization effort for AI visibility, here's what moves the needle more reliably than schema markup:
Third-party citation building
Getting the brand mentioned in authoritative external sources - industry publications, review platforms, directories, practitioner communities - is the highest-impact AI visibility tactic. These mentions are what AI models train on.
Topical depth over breadth
AI models recommend brands that are clearly associated with a specific topic or niche. A brand with deep, consistent coverage of a specific problem set gets recommended for relevant queries more often than generalists.
Entity clarity (About page, Wikipedia, Wikidata)
Having a clear, consistent brand entity definition - well-written About page, Wikipedia entry if warranted, Wikidata record, and consistent brand description across sources - helps AI models accurately identify and describe the brand.
FAQ and conversational content
Content written in a question-and-answer format, directly addressing the kinds of queries AI users ask, is more likely to be surfaced in AI responses. This is where structured data (FAQ schema) complements good content rather than substituting for it.
The bottom line for agency clients
Schema markup is worth implementing correctly - especially FAQ schema for pages targeting question-based queries, and Organization schema for clear entity definition. But it should be one item in a broader AI visibility strategy. Not the centerpiece. This is the part most technical SEO guides get wrong.
The clearest way to understand whether AI visibility is actually improving is to measure it directly: track ChatGPT and Perplexity mention rates over time, and connect content and citation work to changes in those metrics. That's the accountability layer that turns AI visibility from a vague concept into a measurable, improvable channel.
Measure AI visibility directly
See your client's ChatGPT and Perplexity mention rates right now. Free audit, no signup needed.