If you log into ChatGPT, Gemini, or any major AI tool and ask it for a strategy to optimize your website for AI search engines, it will almost certainly spit out the same exact advice:
“Implement structured data and maximize your schema markup.” If you ask it which specific schema to prioritize, it will consistently point you toward FAQ schema (FAQPage) to capture quick answer snippets.
It sounds perfectly logical. AI models run on structured data, so giving them tidy machine-readable data should make them cite your website, right?
Wrong. If you are blindly letting an LLM run your SEO strategy, you are wasting an immense amount of time, energy, and money. Recent industry changes and explosive new data reveal that chasing schema to win AI visibility is a dead end.
Google is Actively Burning Schema to the Ground
The first flaw in the AI’s advice is that it is operating on lagging data. Google has spent the last several years systematically reducing the footprint of rich snippets and structured data features.
The biggest blow came recently when Google officially announced the complete deprecation of FAQ rich results. Google completely removed FAQ search appearances from search results and dropped support for it entirely.
If your agency or an AI tool is telling you to meticulously build out FAQ schema to win Google real estate, they are giving you a playbook for a feature that literally no longer exists. Furthermore, Google recently dropped search support for a handful of other structured data types, including Course Info, Estimated Salary, and Special Announcements.
The writing is on the wall: Google is actively cleaning house and ignoring massive amounts of schema markup.
The Core Data: Schema Does Not Boost AI Citations
Even if Google is dropping schema features on its traditional search pages, does implementing structured data help you get cited inside conversational AI engines like ChatGPT or Perplexity?
An explosive, data-backed study by Ahrefs put this exact assumption to the test—and the results completely shocked the SEO community.
Ahrefs initially analyzed six million URLs and found a massive correlation: pages that were cited by AI engines were three times more likely to have JSON-LD schema than pages that weren’t. On the surface, this looked like proof that schema works.
But correlation is not causation. Highly sophisticated, authoritative websites with incredible content also happen to have clean, technical SEO setups that include schema.
To find the truth, Ahrefs tracked 1,885 pages that newly added schema and matched them against a control group of 4,000 pages to see if adding structured data actually triggered a bump in AI citations.
The result? Adding schema produced absolutely zero meaningful uplift in citations across ChatGPT, Google AI Mode, or AI Overviews. In fact, citations in Google AI Overviews slightly declined.
Spending weeks writing complex code and managing structured schema data is completely irrelevant to an LLM’s live retrieval system.
What Actually Moves the Needle for AI Visibility (LLMO)
If schema is a waste of time, how do you actually optimize your business to get cited by ChatGPT and AI search tools? The same Ahrefs study looked at 1.4 million real-world AI prompts to find out exactly what causes an AI engine to cite one webpage over another.
The answer isn’t technical code. It comes down to one core concept: Semantic Similarity.
When an AI engine processes a query, it searches the web using its own natural machine language. To win the citation, your page content needs to match the exact language and intent the AI is using to look for answers.
Instead of chasing complex backend code, pivot your strategy to these three high-impact tactics:
1. Match the AI’s Search Language
Look at the exact phraseology AI engines use when aggregating data. Ensure that target language is placed exactly where it carries the most weight: your page title, your URL slug, your main H1 header, and right at the beginning of your very first sentence.
2. Prioritize High-Intent Keywords over Informational Queries
A study showed that Google’s AI Overviews drastically reduce organic clicks by 58% on informational queries. Why? Because if someone searches “How old is the Empire State Building,” the AI will cite your page but give the user the answer directly on the screen. The user gets what they want and leaves—meaning you get zero traffic.
Instead, optimize your landing pages for high-intent buyer keywords (e.g., specific services, product pages, or localized solutions) where the user actually needs to click through to complete an action or make a purchase.
3. Build Authentic Entity Authority
AI search models prefer trusted, highly cited real-world sources. Focus your time on things that actually build authority:
- Publishing truly original data, deep case studies, and proprietary benchmarks with sources.
- Creating detailed author profiles and robust “About Us” company pages to prove real-world expertise.
- Earning high-quality backlinks and mentions on trusted, authoritative media platforms.
Final Thoughts
The era of technical “growth hacks” like over-optimizing your schema markup to trick search engines is drawing to a close. AI engines are incredibly smart text-analyzers; they do not need JSON-LD code to understand what your webpage is about.
Stop letting outdated AI prompts dictate a tedious, low-ROI coding strategy. Clean up your basic site navigation, write hyper-relevant content that matches real human (and AI) search intent, and focus heavily on building genuine real-world authority. That is how you win the future of AI search.
This video provided a brilliant technical case study. You can check out a full breakdown of the data on the Ahrefs Audio Log Podcast Episode on YouTube, which dives deep into the exact numbers behind why AI systems entirely bypass JSON-LD code during live data retrieval.
