Structured data is a way of labeling information on your website so that search engines don’t have to guess what it means. On its own, a date on a page is just text. But if it’s labeled properly in the underlying code, it becomes the date of an event or publication that search engines can understand.
For most of the last decade, structured data has been used for search features within Google such as star ratings, recipe photos, and expanding FAQ dropdowns. Over the past two years, Google has removed many of these features, and as a result, one half of the SEO industry has decided structured data is finished while the other half says it’s now the secret to getting mentioned by AI. Both are wrong.
Which Rich Results Google Has Retired, and What This Means for Rankings
In June 2025, Google retired seven structured data types at once: Book Actions, Course Info, Claim Review, Estimated Salary, Learning Video, Special Announcement and Vehicle Listing. The stated reason was that they were not commonly used and no longer added significant value for users. Google also confirmed the change did not affect how pages rank.
As of May 7, 2026, FAQ rich results no longer appear in Google Search, with Search Console reporting removal in June and API support ending in August 2026. There was no blog post. Google added a small banner to existing documentation, and that was the entirety of the announcement. Before that, in November 2024, the sitelinks search box was retired the same quiet way.
This is a noticeable pattern: Search features come and go based on the frequency of people using them, and none of these removals resulted in a larger rankings shift or made the underlying structured data markup invalid.
The only thing Google removed is the display layer, or the decorated version of your SERP listing. Nothing has changed about the code that tells search engines who you are and how your pages relate. Since unused markup causes no problems in Search, there’s no urgency to strip out types that have stopped rendering.
Which Structured Data Types Still Work in 2026
Organization. Organization markup belongs on your homepage, or on a single page describing the company, and not repeated across the site. It has no required properties at all. Google asks you to add whatever is genuinely relevant to your organization. This is the most important type on the list, and we’ll talk about it more below.
Breadcrumbs. Google needs at least two levels, and asks you to represent a typical user path rather than mirroring your URL structure. Including your homepage as the first level is optional, and so is the current page. This feature shows on desktop SERPS only.
Video. Only three properties are actually required: a name, a thumbnail image and an upload date. Be careful with video galleries, since Google works best with one video per page. When it sees several videos at one URL, it may mix up the details, taking a title from one and a thumbnail from another. Give every video its own page and its own thumbnail.
Products. There are two different treatments, and people routinely conflate them. Product snippets are for pages that describe products, such as reviews and roundups. Merchant listings are for pages where someone can actually buy products, and they unlock shipping, returns and sizing detail. You no longer need a Merchant Center account for the second one.
Reviews. This is the one that surprises our clients. Google’s review guidelines say that when a business controls the reviews about itself, its pages are not eligible for star ratings in search results. Your own testimonials page will not produce stars. Neither will embedding a third-party review widget, which the documentation rules out by name. Reviews are still worth collecting; they just earn their visibility through your Business Profile and third-party sites, not through markup on your own pages.
Organization Schema Markup Is the Most Important Type
Organization markup provides your business’s name, logo, description, address, phone number, and social profiles in a machine-readable format. It also houses less obvious properties like industry classification codes, tax and VAT numbers, and business identifiers like a DUNS number, which don’t get displayed anywhere. Google uses them behind the scenes to work out that your company as you describe it and some other record of your company are the same organization.
Every other schema type above feeds into a specific feature that Google could retire at any time. This one makes your business legible, and none of it depends on a display format surviving.
The goal of organization structured data is to be robust, not maximal. Every property that is true and relevant should be filled out, but not every field needs to be treated as a form to complete.
There’s a common piece of advice that says more markup always helps. Google does state that completeness is one of the most important ranking signals, but only for a narrow category called enriched search results, covering three types: Job Posting, Recipe and Event. People tend to stretch this into a general rule. For Organization markup, the documented benefit is being understood correctly, not ranking higher.
Three issues to avoid when updating Organization markup:
- Adding properties that don’t fit – a retailer greatly benefits from shipping and return policy fields, while a law firm would not.
- Guessing at values – inputting an identifier you don’t actually hold or an industry code that’s only roughly correct is worse than leaving it out.
- Adding data nobody will maintain – employee counts and number of active locations are in constant flux. They may be correct at the time of input, but can quickly become inaccurate if not updated regularly.
If you’re unsure which properties to include in your Organization markup, as yourself: Can I actively ensure that every data field is up to date today, tomorrow, and even next year?
Does Structured Data Help AI Search? What the Platforms Say
In a nutshell: They say very little. Google states that there’s no special structured data you need to add to appear in its AI features, and its generative AI guidance repeats that structured data is not required, while still recommending it for rich results. To appear in AI Overviews and AI Mode, a page just must be indexed and eligible to show with a snippet.
OpenAI’s published documentation doesn’t mention schema markup at all. The closest thing to a platform endorsement comes from Microsoft, where a search and AI executive wrote in February 2026 that AI agents browsing the web behave like retrievers, drawn to “structured, verifiable, and applicable content”. But that statement says structured content, not structured data. It may mean well-organized pages rather than schema markup, and stretching it into a claim about schema is exactly how the search industry manufactures certainty it has not earned.
You can now measure some of this. Google reports AI feature performance in Search Console, and Microsoft launched AI citation reporting in Bing Webmaster Tools in 2026, showing how often your content is cited and what share of citations you hold. Microsoft describes that share metric as observational rather than a ranking system, which is a useful reminder of what these numbers are and are not.
So the honest answer on AI visibility: Structured data is not the deciding factor. It’s the layer that makes sure an engine parsing your site understands what it found.
Bing’s Structured Data Guidelines Differ from Google’s
This matters more than it used to, because Microsoft says its grounding technology sits behind nearly every major AI assistant on the market. Even if that’s a marketing-driven exaggeration, it probably means Bing is more than just a rounding error in any good AI strategy.
Bing’s own markup documentation treats markup formats neutrally, saying its crawlers do not prefer one specification over another, whereas Google recommends a single format. It still lists Claim Review as supported, which is one of the seven types Google retired in 2025. It documents markup for people, which Google has no equivalent feature for. And it describes using your annotations to supplement and validate its own data sources, which mirrors how Google treats entities.
It also validates strictly and silently. If a value is the wrong type, or an event has no date, Bing ignores the annotation rather than warning you.
Bing’s markup overview page looks old, referencing formats and conventions that predate its own 2018 support announcements. FAQ markup is not listed anywhere on it, which means the common claim that Bing still rewards FAQ schema is not something Microsoft documents. Treat the documentation as what Bing has written down rather than a complete picture of what Bing does today.
So, Is Schema Markup Still Worth It?
Two things are true at once, and most of the current commentary manages to cover only one of them.
Structured data still matters, and it matters most where it never shows up on a search results page. Google spent two years retiring display features while leaving the entity layer alone, and Bing says outright that it uses your markup to validate its understanding of your business. Make sure your Organization markup is accurate, hold the remaining types to their actual requirements, and most of your markup work will be done.
Structured data is not an AI visibility tactic, and no platform has documented it as one. Fabrice Canel, a principal product manager at Bing, stated verbally in March 2025 that schema markup helps Microsoft’s LLMs understand content for Copilot, but there’s no written evidence that this or anything beyond this is true for any platform. If you’re not sure whether your Organization markup is accurate, or whether the markup you’re using still earns anything, we recommend an that covers structured data and overall organic visibility. Razor Rank’s technical SEO audits include checking structured data against what each platform currently documents, which is a shorter list than most schema advice assumes. And if someone has recently pitched you schema as the way into AI Overviews or ChatGPT citations, ask them which documentation says so. Better yet, contact our team and we’ll give you a second opinion before you sign anything.
Razor Rank is a full-service digital marketing agency specializing in SEO, paid media, CRO, and web. We help businesses grow through data-driven strategy and measurable results.
