Google's FAQ Rich Results Sunset: What Schema Markup Looks Like for Small Business AI Search in 2026

What Exactly Did Google Change About FAQ Rich Results?

Google's deprecation comes in three steps, all spelled out in the official FAQPage documentation. The first step has already happened. As of May 7, 2026, FAQ rich results stopped appearing in Google Search. The second step removes the FAQ search appearance, the rich result report, and FAQ support in the Rich Results Test in June 2026. The third step removes FAQ support from the Search Console API in August 2026, which Google explicitly framed as a window for engineering teams to update API integrations.

Search Engine Land's coverage of the announcement quotes Google directly: “FAQ rich results are no longer appearing in Google Search. We will be dropping the FAQ search appearance, rich result report, and support in the Rich results test in June 2026.” Google's documentation does not require site owners to remove FAQPage markup, and it does not recommend an alternative schema type or migration path. The change is about what Google will display, not about what schema is allowed.

It also helps to be honest about who actually had FAQ rich results in the first place. Google restricted FAQ rich results in August 2023 to “well-known, authoritative websites that are government-focused or health-focused,” language that is still on the official FAQPage page. For the vast majority of small and mid-sized businesses — HVAC contractors, dental practices, regional law firms, e-commerce stores, B2B SaaS startups — FAQ rich results have effectively been gone since late 2023. The May 2026 sunset is the formal end of a feature that, for most of you reading this, has already been off-limits for almost three years. That changes the calibration of how much CTR you should expect to lose.

What is genuinely new is the tooling rollback. The Rich Results Test, the Search Console enhancement reports for FAQ, and the API endpoints that downstream SEO tools depend on are all being removed in the next few months. Anything you have built — a monitoring dashboard, a reporting template, a CI test that pings the Rich Results Test on every deploy — needs to be reviewed and updated.

Why Does FAQ Schema Still Matter If the Rich Results Are Gone?

This is the part most “FAQ schema is dead” takes get wrong. Rich results are one outcome of structured data; they are not the only one. The same FAQPage block that no longer earns a Google SERP feature still does several other things, and several of them have become more valuable since AI search took off.

The first is non-Google search and non-Google features. Schema.org defines FAQPage as a generic web page type for “frequently asked questions,” and it is parsed by Bing, DuckDuckGo, voice assistants, and an expanding set of AI ingestion pipelines. Bing has its own FAQ display logic, voice search assistants pull FAQ pairs into spoken answers, and several AI search tools use schema as one of many signals when deciding which content blocks to ingest. Removing your markup costs you those.

The second is machine-readable Q&A that AI systems can use as a citation anchor. The recent AirOps study covered by Search Engine Land examined 16,851 unique queries across 353,799 pages and found that pages with JSON-LD markup were cited 38.5% of the time versus 32% for pages without it. That same study found pages in the top search position were cited 58.4% of the time versus 14.2% for position ten — meaning ranking still matters more than schema — but the schema delta is real and consistent. Our own AEO playbook for small businesses calls this out as one of the few free, mechanical wins still available.

The third is local and vertical Q&A retrieval. AI assistants increasingly answer questions like “what does HVAC tune-up cost in Fort Wayne” by stitching together fragments of pages — and a structured Q&A pair is much easier to extract cleanly than a buried paragraph. Our post on localized FAQ pages for Fort Wayne businesses goes deeper into how to write those pairs so they survive that extraction.

The fourth, less obvious benefit is internal hygiene. FAQ schema forces you to write content as discrete question-answer pairs, which tends to produce shorter, sharper, more answer-shaped paragraphs. That formatting helps human readers and AI ingestion alike, regardless of whether Google ever surfaces a rich result for it.

Should You Keep, Prune, or Convert Your Existing FAQ Schema?

The right answer depends on what each FAQ block was doing for you. Here is the decision matrix we walk clients through, and the rough percentage of pages we end up putting into each bucket on a typical small business audit.

Those percentages are our internal averages from auditing roughly two dozen small business sites in the last year. They are not a benchmark, and yours will differ. But the framework matters more than the numbers: the post-sunset question is no longer “does this earn a rich result?” — it is “does this section earn its place on the page?”

A few rules of thumb when applying it. If a FAQ entry is answering a question that genuinely repeats in your sales conversations, keep it. If you wrote a FAQ to capture a long-tail keyword that gets meaningful search volume on its own, that question is probably stronger as a dedicated page. If a FAQ entry has been there for three years and you cannot remember why, it is almost always safe to delete. And if you have FAQ pairs where the answer is two short sentences and reads like a definition, you are in the sweet spot ChatGPT citations seem to reward — the AirOps data suggested narrow, focused content beats broad “ultimate guide” pages, and FAQ-style writing is naturally narrow.

There is one place we strongly recommend keeping the schema even when the rich result is gone: pages that already rank in the top three for a transactional query. The cost of leaving valid FAQPage markup in place is roughly zero, and the benefit — a marginal lift in AI citation rate — compounds over thousands of impressions. Strip it only if you are removing the underlying content.

How Do You Repurpose FAQ Content for AI Citations?

The most valuable thing you can do this quarter is not “decide what to do with the schema.” It is “rewrite the FAQ blocks so AI systems will actually pull from them.” Three patterns work best in our experience.

The first is the question-as-H2 rewrite. Instead of nesting twenty FAQs at the bottom of a service page, take the three or four highest-intent questions, promote them to H2s in the body, and write 80–150 word answers immediately under each. AI Overviews and ChatGPT both extract content much more reliably from labeled, top-level sections than from collapsed accordion blocks. We cover the why of this in why your content doesn't appear in AI Overviews, and the same logic applies here.

The second is the data-anchored answer. AI systems increasingly cite content that includes a specific number, a price range, a date, or a comparison. A FAQ that says “Our HVAC tune-up is comprehensive” gets ignored. A FAQ that says “Our HVAC tune-up takes 60–90 minutes, costs between $129 and $179 in 2026, and includes a 21-point inspection list” gets cited because there is something specific to attribute. You do not need to fabricate numbers — you need to surface the ones already in your sales process. We walked through this in detail in our piece on topical authority not being enough for AI search: proprietary data is what differentiates pages that all cover the same topic.

The third is the dedicated Q&A page for high-volume questions. If “how long does a roof inspection take in Fort Wayne” gets searched 50 times a month, that question is too valuable to bury inside a generic services page FAQ. Promote it. Give it its own URL, a clear H1 that mirrors the question, a short direct answer in the first paragraph, and a longer expanded answer below. Use Article schema (which Google still actively supports per the structured data feature gallery) rather than FAQPage. This is the same logic that drove the original “FAQ schema is the hidden AEO powerhouse” thinking we wrote about in our older guide on FAQ schema for AEO — it just happens to apply better as standalone pages now that the rich result is gone.

A note on what not to do. We have seen agencies recommending wholesale deletion of FAQ schema as a defensive move “in case Google penalizes it.” There is no evidence of that — Google's documentation explicitly does not ask site owners to remove the markup. Deleting valid schema costs you AI citation surface area you can't easily get back. If you are unsure, leave it.

What Should You Do in the First 30 Days After the Sunset?

A short, sequenced checklist. Most of this is one to three hours of work for a small business; none of it requires a redesign.

1. Run an inventory of every page on your site that uses FAQPage schema. The fastest way is to crawl your sitemap with Screaming Frog or Sitebulb and filter by JSON-LD type. If you do not have those tools, an “sitename.com FAQ” site search plus your CMS's FAQ block report usually gets you 80% of the way.
2. Pull baseline traffic data for those URLs from Google Search Console. Filter to the last 28 days, export, and save it. You want a “before” snapshot in case you ever need to argue that a traffic drop did or did not correlate with the May 7 change. Most small businesses will see no detectable change because they were not eligible for FAQ rich results to begin with.
3. Apply the keep/prune/convert/delete matrix above to each page. Be honest about the bottom 15%; FAQ filler that exists only to bulk up word count is a small but real drag on AI ingestion quality.
4. Update internal monitoring to remove anything that depends on the Rich Results Test FAQ check, the Search Console FAQ enhancement report, or the FAQ Search Console API (sunset August 2026). If you have CI checks that ping these endpoints, swap them now rather than wait for them to start failing.
5. Pick three to five FAQ-heavy URLs to actually rewrite using the H2-promotion or data-anchor pattern. Do not try to do everything. Pick the pages that already rank or already convert, and improve those first. The AEO uplift from concentrated rewrites consistently beats spreading the same effort across the whole site.

You do not need to do this in a single sprint. We typically recommend spreading it over four to six weeks for small teams, with the inventory and monitoring updates in week one and the rewrites staggered after.

How Does This Change Affect Fort Wayne and Northeast Indiana Small Businesses?

For most Allen County, DeKalb County, and Northeast Indiana small businesses, the immediate traffic impact of this sunset is going to be minimal. The 2023 restriction to government and health sites already cut off most local HVAC, dental, legal, roofing, and home services companies from FAQ rich results almost three years ago. If your Search Console FAQ enhancement report has been quiet since 2023, May 7, 2026 will not look any different.

What does change is the AI citation calculus. Fort Wayne has a structural advantage in AI search that we have written about before: regional service businesses can publish content that no national competitor can — neighborhood-specific pricing, weather-driven service timing for Northeast Indiana winters, named local zoning and permit context. That kind of proprietary, hyper-local content is exactly what AI systems pull from. If your FAQ pages currently have generic answers that could apply to a service business anywhere in the country, the FAQ sunset is your prompt to rewrite them with Auburn, Fort Wayne, or DeKalb County specificity. Replace “How much does HVAC service cost?” with “How much does an HVAC tune-up cost in Fort Wayne in 2026?” and write an answer that uses your actual local price range. That single change moves you out of the “indistinguishable from any other contractor” bucket and into the citable bucket.

We have been seeing this pattern across the Fort Wayne service businesses we work with: localizing FAQ answers — even just adding the metro name and a recent year — meaningfully increases the chance that AI assistants name the business specifically when a Fort Wayne resident asks for a recommendation. The schema doesn't earn the rich result anymore, but the localized answer earns the citation, which is a much bigger prize.