Blog
FAQ Schema on Product Pages: Still Worth It in 2026?
Google stopped showing FAQ rich results in May 2026. What that changes for product pages, what to do with existing markup, and why the FAQs themselves matter more for AI answers than the schema ever did.
By Alberto Barberis, Founder and CEO, AndromedAI · Updated October 2026
The short answer
FAQ schema (FAQPage structured data) no longer produces rich results in Google Search: the feature ended for all sites on May 7, 2026. The markup is harmless and can stay where it mirrors visible text. The value now sits in the content: product-specific questions answered on the page and in Merchant Center, where AI assistants can read and quote them.
FAQ schema: the short answer
For Google rich results, FAQ schema is finished. For AI shopping answers, well-written product FAQs still matter, and they matter as visible text first.
- 01
FAQ rich results are gone from Google Search for every site since May 7, 2026.
- 02
Existing
FAQPagemarkup does no harm, so there is no reason to rip it out in a hurry. - 03
Adding schema alone barely moved AI citations in the largest study so far.
- 04
The questions and answers themselves still earn their place on a product page, in visible text and in your Merchant Center feed.
What changed for FAQ rich results in 2026
Google ended FAQ rich results for every site on May 7, 2026, then removed the reports, the Rich Results Test support and the API data over the following three months.
Google trimmed FAQ rich results in two steps. In August 2023 it restricted them to "well-known, authoritative government and health websites", as Search Engine Journal recalls. Most stores lost the accordion under their listing back then.
The second step finished the job. Google's documentation changelog logged the deprecation on May 8, 2026, with the line "This feature will no longer appear in Google Search starting May 7, 2026", and removed the FAQ documentation on June 15.
| Date | What changed |
|---|---|
| August 2023 | FAQ rich results limited to authoritative government and health sites |
| May 7, 2026 | FAQ rich results stop appearing in Google Search for all sites |
| June 2026 | FAQ search appearance filter, rich result report and Rich Results Test support removed |
| August 2026 | FAQ support removed from the Search Console API |
Google's notice says the markup can stay without causing problems. FAQPage is still a valid schema.org type, so validators that check the vocabulary will keep accepting it.
Does FAQ schema help AI answers?
The markup on its own, barely. The visible answers, yes: AI assistants quote plain sentences that answer narrow product questions, wherever those sentences sit on the page.
Two years of conference talks claimed FAQ markup is how you get quoted by AI. The best data we have says otherwise.
Ahrefs tracked 1,885 pages that added JSON-LD between August 2025 and March 2026 and compared them with 4,000 matched control pages. After the change, citations moved +2.2% in ChatGPT and +2.4% in Google AI Mode, neither distinguishable from zero, and -4.6% in AI Overviews. The study pooled schema types, so it can't isolate FAQ markup, but it shows that schema on its own isn't a lever for pages already visible to AI.
Google's AI features guidance states there is "no special schema.org structured data that you need to add" and asks for important content "in textual form". Microsoft is a little warmer: Bing's Fabrice Canel confirmed that schema markup helps its LLMs understand content.
What does reach the answer is the visible sentence. A shopper asks something narrow, the assistant looks for a page that answers it plainly, and it repeats that answer.
Cup holder fit and dishwasher safety rarely appear in a title or a spec table. They sit in FAQs, and that's why we keep writing them. Shoppers who arrive this way are worth the effort: Adobe measured AI-referred visitors converting 60% better than other traffic to US retail sites in July 2026.
Where product FAQs should live now
On the product page in crawlable HTML first, then in your Merchant Center feed. FAQPage markup is a mirror of that content, never a substitute for it.
In 2026 a product FAQ has four possible homes, and they don't carry equal weight.
| Where the FAQ lives | Who reads it | Verdict |
|---|---|---|
| Visible text on the product page | Shoppers, Google, AI crawlers and agents | Required. Everything else depends on it |
FAQPage JSON-LD | Validators, Bing, some parsers | Optional. Keep it if it matches the visible text |
Merchant Center question_and_answer | Google AI Mode and Gemini | Worth adding for your top products |
| OpenAI product feed | ChatGPT shopping | No Q&A field, so the page has to carry it |
The Merchant Center attribute is the newer development. Google introduced six conversational attributes at Google Marketing Live 2026 for "AI-powered shopping experiences, such as AI Mode and Gemini", Productsup reports. The question_and_answer attribute takes up to 30 pairs per product, 1,000 characters per question or answer and 10,000 characters in total. Google asks you to leave prices and dates out, skip keyword stuffing and avoid repeating what is already in the title or description.
OpenAI went the other way. Its product feed reference says "question-and-answer lists are not part of this discovery contract." For ChatGPT shopping, your FAQ only counts if it is on a crawlable page.
The markup mistakes we still see
The most common one is a Shopify or WooCommerce app that injects the same shipping and returns FAQPage block on every product. One store FAQ copied onto thousands of URLs was weak before May 2026 and is pointless now.
Second, answers fetched by JavaScript only when the accordion opens. Google's structured data policies ban marking up content "not visible to readers of the page", and an AI crawler that doesn't run your scripts won't see those answers at all. Put the answers in the initial HTML.
Third, the wrong type. If customers post questions and other customers answer, that's a Q&A widget, and Google's QAPage documentation describes it separately from a merchant-written FAQ. Don't relabel review Q&A as FAQPage.
How to write product FAQs AI assistants can quote
Use real customer questions, keep a handful per product, put the fact in the first sentence and keep every answer consistent with your product attributes.
A product FAQ works for AI answers when each answer states a fact the rest of the page doesn't already say, in the first sentence, about that product only.
- 1
Mine the real questions
Support tickets, return reasons, reviews and on-site search beat any keyword tool here.
- 2
Keep 4 to 8 per product
Thirty weak questions dilute the five good ones.
- 3
Lead with the fact
"Yes, the lid is dishwasher safe on the top rack."
- 4
Match your attributes
If the FAQ says 100% merino, the
materialfield and the spec table must say it too. Conflicts make agents distrust the page. - 5
Leave offers out
No prices, discounts or delivery dates in answers. Merchant Center asks you to keep those in offer attributes.
- 6
Mirror the best ones
Send the strongest pairs to Merchant Center as
question_and_answer, and keepFAQPagemarkup only where it mirrors the visible text.
faq
BeforeQ: Shipping times? A: See our shipping page.
AfterQ: Will it fit a car cup holder? A: Yes, the base is 7 cm wide and fits standard cup holders.;; Q: Can it go in the dishwasher? A: Yes, lid and body are dishwasher safe on the top rack.
description
BeforeOur best bottle for every adventure.
AfterDouble-wall 18/8 stainless steel bottle, 24 oz (710 ml). Keeps drinks cold for 24 hours and hot for 12. Leak-proof screw lid with a carry loop.
attributes
Beforecolor: blue
Aftermaterial: 18/8 stainless steel; capacity: 24 oz; color: harbor blue; dishwasher_safe: yes; base_diameter: 7 cm
The "before" FAQ is the generic kind we find on most catalogs we audit. It answers a store question and says nothing about the bottle. For the description side of the same fix, see how to write product descriptions, and run the finished page through our product page checklist for AI readiness to catch the fields the FAQ depends on.
A product FAQ earns its place when it answers something the title, description and spec table don't.
Should you keep, remove or add FAQ schema?
Keep product-specific markup that matches the page, remove sitewide boilerplate, and put your effort into the content and the feed instead of the JSON-LD.
This is the order we work in when a team asks what to do with markup they already have.
- Audit before you delete
Crawl a sample of product detail pages (PDPs) and list which
FAQPageblocks are product specific and which are boilerplate - Remove sitewide boilerplate
Shipping and returns questions belong on one help page, with a link from the PDP
- Keep matching markup
Product-specific FAQ markup that mirrors the page costs nothing and still describes the page for other parsers
- Fix rendering
Answers must be in the initial HTML, not fetched when the accordion opens
- Update reporting
Drop FAQ rich result KPIs and annotate the August 2026 API change in dashboards
- Add the feed version
Submit
question_and_answerfor bestsellers first, in TSV to avoid escaping trouble with commas and colons
We'd disagree with one popular reaction. Deleting product FAQs because the rich result is gone throws away the content AI assistants quote most readily. Baymard rates 51% of ecommerce sites as "mediocre" or worse on product page UX, and missing answers to basic questions are part of that.
For the full markup picture (Product, Offer, ProductGroup, reviews), see our guide to structured data for ecommerce. For how Google's shopping assistant uses product data, read Google AI Mode shopping.
How AndromedAI helps
AndromedAI writes product-specific FAQs from your catalog data and publishes them to your store and to Merchant Center.
The Optimizer writes product-specific FAQs from your catalog data along with titles, descriptions and bullets, using your Brand Kit so the answers sound like you. It extracts the attributes those answers rely on, so the FAQ and the material or size fields agree. It also generates Merchant Center conversational attributes and publishes to Shopify, Google Merchant Center and the other platforms on our integrations page. This is one piece of Agentic Commerce Optimization (ACO): making product data readable for search engines, ads and AI agents at once.
clicks from AI chats
Bomboogieto first sales from ChatGPT
Matassaorganic traffic
SemprefarmaciaFAQ
Not for Google rich results, which ended for all sites on May 7, 2026. FAQPage is still a valid schema.org type, and the markup does no harm where it mirrors questions and answers visible on the page. What still pays off is the FAQ content itself, because AI assistants read and quote the visible answers.
No. Google stopped showing FAQ rich results for all sites on May 7, 2026. It then removed the FAQ search appearance filter, the rich result report and Rich Results Test support in June 2026, and dropped FAQ data from the Search Console API in August 2026.
Only where it is boilerplate, such as the same shipping and returns questions injected on every product by an app. Product-specific FAQ markup that matches the visible answers can stay, since it costs nothing. Move generic store questions to one help page and link to it from your product pages.
Not on its own. An Ahrefs study of 1,885 pages that added JSON-LD found no meaningful citation gain in ChatGPT or Google AI Mode, and a small drop in AI Overviews. What helps is a clear answer in visible text on the page, which assistants can find and repeat.
Usually 4 to 8 product-specific questions about fit, care, compatibility, contents or use. Merchant Center accepts up to 30 question_and_answer pairs per product, but a handful of strong, factual answers is more useful than a long list. Thirty weak questions dilute the few that shoppers actually ask.
It is an optional conversational attribute Google introduced in 2026 for AI shopping experiences such as AI Mode and Gemini. Each product can carry up to 30 question and answer pairs, with up to 1,000 characters per question or answer. Google asks merchants to leave prices and dates out of the answers.
Glossary
- FAQ schema
- Structured data using the schema.org FAQPage type to mark up a list of questions with one answer each, written by the site owner.
- FAQ rich result
- The expandable question list Google once showed under a search result for pages with FAQPage markup. Discontinued for all sites on May 7, 2026.
- QAPage
- The schema.org type for a page focused on one question where users can submit answers, such as a forum thread or a product support question.
- question_and_answer
- An optional Merchant Center attribute for product FAQs, with up to 30 pairs per product, used in conversational experiences such as AI Mode.
- Conversational attributes
- Six optional Merchant Center attributes introduced in 2026 to give AI shopping experiences richer product context.
Keep reading
The full markup guide this article belongs to
Read nextGoogle AI Mode Shopping: How Products Get Picked in AI Mode and GeminiHow Google's assistant reads your product data
Read nextProduct Page Optimization: The Complete Guide to PDPs That Rank, Convert and Get Recommended by AIWhere FAQs fit on a high-performing PDP
Read nextProduct Page Checklist for AI ReadinessCheck every field your product FAQs rely on
Sources (13)
- Search Engine Journal: Google drops FAQ rich results
- Google Search Central: Latest documentation updates
- Google Search Central: AI features and your website
- Google Search Central: General structured data guidelines
- Google Search Central: Q&A (QAPage) structured data
- Google Merchant Center Help: Question and answer [question_and_answer]
- Productsup: Google introduces 6 conversational attributes in Merchant Center
- OpenAI Commerce: Product feed reference
- Ahrefs: We tracked 1,885 pages adding schema
- Search Engine Land: Microsoft Bing and Copilot use schema for its LLMs
- Adobe Digital Insights: AI Traffic Trends Report, August 2026
- Baymard Institute: Product Page UX research
- Schema.org: FAQPage
Featured in
