On 7 May 2026, Google quietly switched off something thousands of Singapore businesses had spent years building their FAQ pages around. FAQ rich results, the expandable question-and-answer snippets that once dotted the search results page, stopped appearing in Google Search altogether. No fanfare, no blog post, just a line added to a documentation page.
For anyone who has spent the past two years being told that FAQ schema is the fastest way into an AI Overview or a ChatGPT answer, the timing raises an obvious question. If Google no longer wants FAQ markup cluttering its results, does the format still carry any weight with the AI platforms now doing a growing share of the answering.
What Actually Changed on 7 May 2026
Google’s FAQ structured data documentation now states plainly that FAQ rich results are no longer appearing in Google Search, and the wind-down continues on a published schedule. The FAQ filter and rich result report leave Search Console in June, and support for FAQ data in the Search Console API ends in August. Sites can leave the FAQPage markup in place without penalty. It simply stops doing anything inside Google’s own results pages.
Why Google Walked Away From FAQ Rich Results
The removal did not arrive out of nowhere. Google had already restricted FAQ rich results to a small list of authoritative government and health sites back in 2023, after the format was widely gamed on pages that never answered a genuine user question. Industry commentators reading the final removal have offered different theories. Some see it as Google reclaiming the answer layer of its results page for itself, now that AI Overviews already summarise most FAQ-style queries directly. Others argue it is simply Google tidying up a feature it had already gutted three years earlier.
What nobody at Google has said, in any official capacity, is that the format itself stopped carrying weight with AI-generated answers. The two claims, rich result eligibility on Google Search and usefulness to an AI model reading a page, were never the same claim to begin with.
The Markup and the Writing Are Not the Same Thing
This is where a lot of the confusion sets in. FAQPage schema is a piece of markup, a block of JSON-LD code that tells a crawler a particular section of a page contains questions and answers. FAQ content is the writing itself, the actual question phrased the way a customer would ask it, followed by a direct, specific answer. Google’s change affects the first of those and says nothing conclusive about the second.
It helps to place FAQ markup within the wider toolkit it always belonged to. Organisation markup, product data, article metadata and the broader entity authority signals AI platforms draw on to work out who a business is and whether to trust it have not gone anywhere, and most AI retrieval systems were never leaning on FAQPage tags specifically in the first place. Readers who want the fuller picture of how that structured data layer works can find it in our entity authority playbook, which covers the schema types that do most of the heavy lifting for AI visibility.
Large language models mostly work from the visible text of a page when they retrieve and summarise content, not from the markup sitting underneath it. That is precisely why the writing itself deserves more attention than the code wrapped around it.
What the Data Says About Q&A Content and AI Citations
A large-scale study comparing AI-cited pages against top-ranking pages, run by Semrush across roughly 305,000 AI-cited pages and more than 920,000 pages ranking in Google’s top twenty, found that several content qualities showed up far more often on the cited pages. Q&A formatting was one of the strongest of them.
- Clarity and summarisation, meaning a short, direct statement of the key point near the top of the page, appeared on cited pages 32.8% more often.
- Strong experience and expertise signals, such as named authorship and clear sourcing, showed a 30.6% difference.
- Content written in a genuine question-and-answer format showed a 25.5% difference.
- Clear section structure, using headings and lists to break content into retrievable chunks, showed a 22.9% difference.
- Visibly structured data within the text itself, separate from any schema markup, showed a 21.6% difference.
These are associations drawn from a single snapshot in time, not guaranteed causes, and the researchers themselves note that citation patterns shift quickly as platforms update their retrieval methods. Even so, the direction is consistent with what several platform-specific studies on this blog have already found about how generative engines select and excerpt sources.
Writing FAQ Content AI Platforms Actually Cite
None of this calls for ripping FAQ sections off a website. It calls for writing them for the reader an AI model is summarising on behalf of, rather than for a rich snippet that no longer exists, and a few habits tend to separate the FAQ sections that keep getting pulled into answers from the ones that quietly disappear once Google stops rewarding the markup around them.
- Phrase each question the way a real customer types it into a search bar or asks a chatbot, not the way a marketing brief would word it.
- Answer in the first sentence, then add supporting detail afterwards, so the answer still makes sense if an AI model only lifts the opening line.
- Keep each answer specific to one question. A paragraph trying to cover three related points at once is harder for a model to excerpt cleanly.
- Update FAQ content when the underlying facts change, since several of the AI platforms covered on this blog weight recency heavily when choosing what to cite.
Where FAQ Content Fits in the Wider AI Search Picture
The removal of FAQ rich results is a reminder that the tactics built around pleasing a single search engine’s results page were always a narrower bet than the content strategy sitting underneath them. The rich-result markup was only ever a wrapper. The underlying habit of answering real questions clearly, in language a customer and a language model can both follow, is the part that was doing the actual work.
Businesses that use this moment to tidy up thin, promotional FAQ sections, instead of mourning a lost rich snippet, are likely to come out ahead as AI search keeps growing its share of how people find Singapore businesses. If you would like a second opinion on where your own content stands, feel free to get in touch with our team.