The truth SEO agencies don't want to hear: the scheme is not required for LLMs
SEO Laboratory 
2025/10/08 
Kastytis 
In the world of SEO, few topics have caused as much controversy as the question of whether "scheme" notation helps content to stand out in artificial intelligence search results. Many SEO influencers and agencies have long argued that structured data is the secret to becoming "visible" to AI systems such as ChatGPT, Gemini or Perplexity.
But there is now growing evidence that this idea is wrong.
New research and field tests show that The scheme does not use large language models (LLM) for generating responses. In fact, these models are not even able to "read" it properly.
🧪 What does the research show?
Portal Search Engine Roundtable summarised it as follows:
"Structured data does not help increase visibility in AI search."
Research by renowned SEO experts Mark Williams-Cook and Julio C, showed the same thing: LLMs are simply doesn't understand schema markup the way Google's algorithms do.
How LLMs see the scheme
LLMs work by tokenisation, breaking text into small units called "tokens". When schema code is broken down, it loses its meaning. Instead of recognising "@type: organization" as structured information, the AI sees only individual words "type" and "organisation".
This causes the structural data (schema) to lose all the context that helps search engines. For LLM, the schema looks like random text or even coderather than meaningful information.
Julio C carried out a test with two almost identical product pages:
- One was visible text and structured data
- Next, only the schemewithout any visible content
When ChatGPT and Gemini were asked to extract product information, they were only able to do so from a page with visible text. The diagram without text went unnoticed.
His conclusion was clear:
"If your agency says that the scheme helps you to be mentioned in AI channels, they are just wrong or misleading."
📊 Correlation is not a reason
Some SEO experts have noticed that websites with structured data appear more often in AI results. This is true, but it is a correlation, not a cause.
Structured data "Schema" helps for traditional Google indexing, which is why pages with it often have higher rankings. And higher rankings increase the chance of being cited in AI responses. But the scheme itself is not the reason why LLMs use your content.
💪 What really helps you get into AI search?
If "schema" is not the magic path to AI visibility, what works?
Here are some practical SEO tips that actually increase the chances of being mentioned AI search engines and LLMs Answers:
1. Strengthen your brand pages
Yours home page and "About us" page you need to be clear about who you are and why your company is worthy of attention. Include:
- For real feedback and recommendations
- Prices and contact details
- Matoma FAQ (FAQ)and not hidden under tabs
Study "SearchPilot" showed that pages with visible text were 12% more organic flow than those where the text was hidden.
2. Control the content of the comparison
LLMs often compare products and services. Create comparison or alternatives pagesthat present your brand as a better choice. Use keywords like "alternative [competitor]" via "[your brand] vs [competitor]".
3. Use press releases strategically
Press releases still influence AI systems. Even low-cost messages help ChatGPT and Perplexity link your brand to relevant topics. Just make sure the text is unique.
4. Write readable, human-centred content
LLMs understand better simple, clear text. Don't overload the page with complex codes or hidden data. The more naturally written the content, the better it is perceived by both readers and AI.
💡 The essential message
"Schema" still useful for traditional SEO, it helps to show extended results, index your content and make it more search engine friendly. But when it comes to AI discovery, chat systems and LLM guidance, scheme just not working.
If you're spending a lot of time or money on a niche scheme, you'd better stop. Instead, pay attention:
- For strong brand content
- For clear, comprehensive pages
- For quality content and natural keywords
- Consistent communication through press releases and quotes
AI in search clarity and authority always wins against code and complexity.
📌 In conclusion
Structural data "Schema" tagging used to seem like an SEO miracle. But in 2025, the truth is clear: LLMs ignore your scheme data. They value content, reliability and visibility.
So before you pay for "AI schema optimisation", ask for proof. Or better still, invest in better content. Because that's what both humans and AI will read.
Vilnius, Lithuania