Blog
Does FAQ schema still help now that AI answers exist?
By Bhavesh Barot, Updated

If your site already builds FAQ schema (structured data, a few lines of code that label the questions and answers on a page) from the FAQ visitors can see, leave it on, but don't add it by hand to get into AI answers. Google says its AI Overviews, the AI-written summaries it adds to some results pages, need no special schema.
We checked the pages those AI Overviews cited on 15 November 2026. Of the 111 cited pages we could read, 30 had FAQ schema in the HTML their servers sent and 81 did not. That is a count, not a cause. It cannot show whether the schema helped any page get cited.
FAQ schema no longer changes what searchers see in Google either. FAQ rich results, the questions and answers Google built from it and showed with a page's listing, were limited to "well-known, authoritative government and health websites" from August 2023. On 7 May 2026 they stopped for every site.
What did we count?
On 15 November 2026, between 02:48 and 02:56 UTC, we ran 61 Google searches through DataForSEO, a data service that returns Google's results page as Google served it. They included "how to choose a chartered accountant", with Google's location set to India, at 02:48 UTC, and "best employment lawyer in chicago", set to Chicago, at 02:49 UTC. The rest were 'best ...' and 'near me' searches set to Bengaluru, other 'how to choose' and fee and process questions in India, 'best CA firm' (chartered accountant) in seven Indian cities, questions US buyers ask lawyers, dentists or roofers, and 'best lawyer' in nine more US cities. Of the 61, 40 showed an AI Overview, both named searches among them.
At 02:57 UTC, we fetched each cited page whose address we could see with a plain request that identified itself as a browser. No page scripts ran. A page counted as having FAQ schema if its HTML carried a JSON-LD block (structured data in its own script block) of type FAQPage, or FAQPage microdata (labels written into the page's HTML tags).
| What we counted | Number |
|---|---|
| Google searches run | 61 |
| Searches that showed an AI Overview | 40 |
| Different links those AI Overviews cited as sources | 282 |
| Google redirect links that hid the page (left out) | 141 |
| Links that showed their real page | 141 |
| Pages that blocked our request, or where the request failed | 30 |
| Pages we could read | 111 |
| Pages we could read with FAQ schema in their HTML | 30 of 111 (27%) |
| Pages we could read with any JSON-LD in their HTML | 85 of 111 (77%) |
So 81 of the 111 pages had no FAQ schema in the HTML their servers sent, while 85 carried at least one JSON-LD block of some kind.
In all, 171 of the 282 links (61%) never reached the count. Google returned 141 as its own redirect links, which hide the page they lead to, so we left them out rather than guess. Another 30 pages blocked the request or it failed, and we did not try to get round a block.
What doesn't the count show?
It shows that pages get cited in AI Overviews without FAQ schema in their HTML, which fits what Google says. Whether FAQ schema helps or hurts, it cannot tell you:
- Cause. We cannot say the schema made Google cite any of the 30 pages that had it.
- A baseline. We counted only pages Google cited, so we can't say whether 27% is high or low.
- Quality. We checked that FAQPage markup was there, not that it was valid or matched the page.
- Scripts. Schema added by JavaScript after a page loads would not show in the HTML we read.
- Missing pages. We don't know what the 141 hidden links or the 30 unreadable pages carried.
- One day, one engine. These are one day's searches, in Google's AI Overviews only. Another day could cite different pages, and we checked no other AI assistant.
Does Google still use FAQ schema?
Not for anything a searcher sees:
| Date | What Google did |
|---|---|
| 8 August 2023 | Announced that FAQ rich results would show only for "well-known, authoritative government and health websites", and no longer regularly for other sites |
| 7 May 2026 | Stopped showing FAQ rich results in Google Search, according to its changelog entry of 8 May 2026 |
| 15 June 2026 | Removed its FAQ rich result documentation, because the feature "is no longer shown in Google Search results" |
On 15 November 2026, the old documentation address forwarded to that changelog, and Google's list of supported structured data features (last updated 16 October 2026) named 25 features, none of them FAQ.
For its AI answers, Google's page on AI features and your website (last updated 10 December 2025) says there's "no special schema.org structured data that you need to add". Its generative AI guide (last updated 10 July 2026) lists "Overfocusing on structured data" among things you can ignore, next to llms.txt files. It still recommends structured data for rich results, which no longer include FAQ. Google does ask, on the AI features page, that any structured data you keep matches the visible text.
Should you keep, add or remove FAQ schema?
Keep it where it costs you nothing and still matches the page. Don't put new work into it.
- Your site builds it from the FAQ on the page. Leave it on. In 2023 Google said there was "no need to proactively remove it", as structured data it doesn't use causes no problems for Search.
- It means typing every answer twice. We would skip it. It earns no Google rich result, and Google's AI features don't need it.
- The code and the page no longer match. A plugin or an old template can leave stale questions in the code. Fix or delete them, because Google's rule is not to mark up what visitors can't see.
- Someone sells it as the way into AI answers. Ask for their evidence. In our check, 81 of the 111 cited pages we could read had no FAQ schema in their HTML. Google's guide also says work aimed at AI answers is "still SEO"; more on that in our post on answer engine optimisation.
Our blog builds each post's FAQ schema in code from the same list of questions shown at the end of the post, this one included, so nothing is typed twice. On our post about what your AI visibility score means, all 12 FAQ entries in the code matched the page word for word when we checked on 15 November 2026. We keep it because it costs nothing once built, not because we expect it to get a post cited.
What should the page say instead?
The answers themselves. Google's guide says its AI answers are written from "the specific information" on the pages they retrieve, so put buyers' real questions on the page, with plain-text answers that use your own specifics: timelines, the documents you need, who does the work and the areas you cover.
Then see what AI answers say when buyers ask. By hand, ask an AI assistant a buyer's question and note the date, the firms named, their order and the sites linked.
A Seen in Search report does this with a standard method. It writes the buyer questions from what your homepage says you sell, and scores the answers in code with fixed, published weights. One of its five measurements is citations, worth 15% of the score: how often the answers link to your own website. The free report a visitor gets today checks ChatGPT only; the full paid report also covers Google's AI Overviews and Perplexity. The methodology page lists the weights and the size of each report. Reports are private when they finish.
One free report per business. Use your business email. No card.
How do you spot FAQ schema on your own pages?
View a page's source in your browser and search for "FAQPage", which is what our check looked for. Where FAQ schema is present, it holds these parts, shown here for an illustrative made-up firm with one question:
| Part of the markup | What it holds |
|---|---|
"@type": "FAQPage" | Says the page is a list of frequently asked questions |
mainEntity | The list of questions |
"@type": "Question" | One question in that list |
name | The question's wording: "Do you work with startups?" |
acceptedAnswer | The answer to that question |
text | The answer's wording: "Yes. We work with startups from their first year of trading." |
The name and text should be the words a visitor reads. To check the code, paste the page's address into the Schema Markup Validator, which schema.org says checks that markup is used "in widely understood ways", not against any one service's needs.
Questions owners ask about FAQ schema
Google showed these under "People also ask" for searches about FAQ schema and schema markup on 15 November 2026.
What is a FAQ schema and how does it work?
Is FAQ schema still relevant?
Are FAQs beneficial for SEO?
Does schema help with SEO?
Does AI use schema markup?
How to check if a website has schema markup?
How to check if schema is working?
What is the difference between schema and structured data?
Can you give me an example of schema markup?
What is the difference between JSON and JSON-LD?
How to apply schema markup?
One free report per business, private unless you publish it. Use your business email. No card. Questions about the report? Contact us.
Sources
- Google Search Central, Changes to HowTo and FAQ rich results, John Mueller, 8 August 2023, read 15 November 2026.
- Google Search Central, Latest documentation updates, entries "Deprecating the FAQ rich result feature" (8 May 2026) and "Removing documentation for the FAQ rich result feature" (15 June 2026), read 15 November 2026. Google's old FAQ documentation address (developers.google.com/search/docs/appearance/structured-data/faqpage) forwarded to this page when we opened it that day.
- Google Search Central, AI features and your website, last updated 10 December 2025, read 15 November 2026.
- Google Search Central, Optimizing your website for generative AI features on Google Search, added 15 May 2026, last updated 10 July 2026, read 15 November 2026.
- Google Search Central, Introduction to structured data markup in Google Search, last updated 10 December 2025, read 15 November 2026.
- Google Search Central, Structured data markup that Google Search supports, last updated 16 October 2026, read 15 November 2026.
- Schema.org, FAQPage, version 30.1 of 16 September 2026, read 15 November 2026.
- Google, Rich Results Test and its help page, read 15 November 2026.
- Schema.org, Schema Markup Validator and Announcing Schema Markup Validator: validator.schema.org (beta), 21 May 2021, read 15 November 2026.
- Introducing JSON, json.org, and W3C, JSON-LD 1.1, W3C Recommendation, 16 July 2020, both read 15 November 2026.
- Seen in Search, How the score works, read 15 November 2026.
- Seen in Search, What does your AI visibility score mean?, its FAQ schema compared with the visible page, checked 15 November 2026, 03:56 UTC.
- Our own check: FAQ schema on the pages cited by Google's AI Overviews in our 61 searches of 15 November 2026 (found through DataForSEO's Google search data; searches run 02:48 to 02:56 UTC, including "how to choose a chartered accountant", India, and "best employment lawyer in chicago", Chicago), checked 15 November 2026, 02:57 UTC.