If you’ve sat through a marketing pitch in the last year, you’ve probably heard some version of this line: “Add FAQ schema to your pages and you’ll start showing up in AI search.” It’s clean, it sounds technical enough to be true, and it gives you something concrete to buy.
It’s also wrong — and as of this spring, there’s hard data to prove it.
Two things happened in May 2026 that, taken together, should change how you think about structured data and AI visibility. Most of the industry reacted to them by picking one of two overcorrections: “schema is dead” or “schema matters more than ever for AI.” Both are wrong. The truth sits in the middle, and it’s more useful than either hot take.
What actually changed
On May 7, 2026, Google quietly added a deprecation notice to its FAQ structured data documentation. FAQ rich results — the expandable question-and-answer dropdowns that used to appear under your search listing — stopped showing up in Google Search. The supporting tools follow in stages: the Search Console report and Rich Results Test lose FAQ support in June 2026, and the Search Console API drops it in August 2026.
For most businesses, this was less dramatic than it sounds. Back in 2023, Google had already restricted FAQ rich results to well-known government and health sites, so the majority of commercial publishers lost that SERP feature years ago. The May 2026 change just finished the job.
Here’s the distinction that got blurred in all the panic: the rich result is gone, but the markup is not. FAQPage is still a valid schema type. You can leave it on your pages — it won’t throw errors, trigger penalties, or hurt your rankings. What ended was the visible search feature, not the code.
So if the schema still works, the reasonable next question is: does it help you get cited in AI answers instead? That’s the pitch a lot of vendors pivoted to the moment the rich result died. And that’s the claim the data now undercuts.
Why the “schema = AI visibility” pitch sounds convincing
The pitch leans on a real correlation. When Ahrefs looked across roughly 6 million URLs, pages that got cited by AI were about three times more likely to have JSON-LD schema than pages that weren’t. If you stop there, the conclusion writes itself: add schema, get cited.
But correlation isn’t causation — and this is a textbook case of the two getting confused. Pages that carry schema tend to belong to well-run sites that also invest in good content, earn links, and build authority. The schema isn’t causing the citations. It’s a marker of sites that are doing everything else right.
What the controlled test found
To separate the two, Ahrefs ran a proper experiment. They tracked 1,885 pages that added JSON-LD schema between August 2025 and March 2026, and matched each one against control pages — similar pages, similar citation levels, but no new schema. Then they measured what happened to citations across Google AI Overviews, Google AI Mode, and ChatGPT in the 30 days before and after.
The result: no meaningful lift on any platform. AI Overviews citations actually slipped about 4.6% for the pages that added schema, AI Mode moved up a couple of points, ChatGPT moved up a couple of points — all small enough to be noise. Four separate tests came back with the same flat answer.
There’s a mechanical reason for this. A separate experiment found that the major AI systems, when they fetch a page in real time, pull from the visible HTML — the words a human actually sees — not from the JSON-LD buried in the markup. The schema layer isn’t where the load-bearing work happens.
One honest caveat, because we care about that here: the Ahrefs study looked only at pages that were already heavily cited by AI. It doesn’t prove schema is worthless for a page that has no AI visibility yet. But it does demolish the specific promise being sold — that bolting on FAQ schema is a reliable path into AI answers.
What actually earns AI citations
If it isn’t the markup, what is it? The same signals that have always separated content worth citing from content that isn’t:
- Answer-first content. AI systems favor pages that state the answer plainly and early. One analysis found pages whose headlines and openings directly answered the query got cited noticeably more often than pages that circled the topic before getting to the point.
- Focused, specific answers over sprawling guides. A tight page that fully answers one question often outperforms a 5,000-word everything-guide, because it’s easier to extract a clean answer from.
- Earned authority and entity strength. Being a recognized, trusted source on a topic — with citations and mentions from other credible sites — is what makes an AI comfortable quoting you.
- Visible, well-structured content. Clear headings, real questions with real answers on the page, accurate specifics. This is the part FAQ content still does well — just not because of the schema wrapped around it.
Google’s own guidance backs this up: it says no special schema is required to appear in AI Overviews or AI Mode. Structured data should match what’s visibly on the page — but it isn’t a secret entry code.
So should you rip out your schema?
No. Keep it. Well-formed schema is cheap comprehension infrastructure — it helps machines understand what your content is, supports the rich results that still exist, and strengthens how search engines model your brand as an entity. That’s genuine value.
What you shouldn’t do is buy schema — FAQ or otherwise — as an AI-visibility strategy, or pay a premium for a tool whose whole pitch rests on that claim. If a vendor tells you structured data is the lever that gets you into ChatGPT or AI Overviews, ask them for the controlled data. The best evidence we have right now points the other way.
The bottom line
The “add code, get cited” era of AI optimization was always a shortcut that skipped the hard part. The hard part hasn’t changed: clear answers, real authority, content genuinely worth quoting. Schema supports that work. It can’t replace it.
If someone’s selling you the shortcut, that tells you more about the seller than about your search strategy.
KDGS Works builds SEO and AI-search strategies on evidence, not vendor talking points. If you want an honest read on where your visibility actually stands, get in touch.
