Schema Markup in 2026

Structured data has always attracted a bit of magical thinking — add the right markup and Google will supposedly reward you with rich snippets, higher rankings, or now, citations in ChatGPT and AI Overviews. Some of that is true. A lot of it isn’t, and 2026 has given us some unusually clear evidence of the difference.

Here’s what schema markup is actually doing for you right now, and where it’s cost without benefit.

The clearest example of cargo-culting: FAQ rich results

For years, adding FAQPage schema to a page — marking up a list of questions and answers — was standard SEO advice, chased partly because it could win you a large, eye-catching expandable box in Google’s search results.

On 7 May 2026, Google discontinued FAQ rich results entirely. It’s not a niche restriction; Google quietly removed the search appearance filter, and by August 2026 will retire the related Search Console API support too. Google’s own note on it is blunt:

FAQ structured data can stay in place. The markup won’t cause problems, but it also won’t produce visible results in Google Search.

This is the third act of a story that started in 2023, when Google first restricted HowTo rich results to mobile results for government and health sites only, then progressively narrowed FAQ eligibility before dropping it altogether. If your SEO checklist still tells you to wrap every FAQ section in FAQPage schema for the rich snippet, that checklist is now three years out of date.

The nuance: the schema itself isn’t necessarily wasted effort — some practitioners believe well-structured FAQ content may still help AI systems parse question-and-answer pairs, even without a Google rich result attached. But if the reason you were doing it was the search-results real estate, that reason is gone.

What’s actually confirmed to help

Google AI Overviews. Google has stated directly that structured data gives content an advantage in AI Overviews results. This is one of the few places where there’s an explicit, on-the-record statement from a major platform that schema markup helps.

Microsoft Bing Copilot. Schema markup helps Bing’s systems understand content well enough to use it for citations — again, a case with reasonably direct platform guidance behind it.

Entity clarity, generally. Organization, Person, Product and Article schema — implemented as an interconnected entity graph using @id references and @graph structures — is the markup with the strongest platform guidance behind it across the board. It’s less about winning a specific rich result and more about removing ambiguity: telling any system, human-built search index or AI model, unambiguously who you are, what you sell, and how your pages relate to each other.

What’s unproven or oversold

“Schema gets you cited in ChatGPT/Perplexity.” A December 2024 study looking at citation rates found no meaningful correlation between how much schema markup a site had and how often it got cited by AI answer engines. Sites with comprehensive markup didn’t reliably outperform sites with minimal markup. If a pitch promises schema will get you quoted by AI chatbots, that promise isn’t backed by the evidence so far.

Schema as a substitute for content quality. There’s a laboratory-style finding that language models extract information more accurately from structured prompts than unstructured text — true, but it describes how models process a prompt, not proof that AI search platforms preserve or use your schema markup while crawling the open web. Structured data can’t compensate for thin content or weak topical authority, and treating it that way is exactly the cargo-culting trap: doing the ritual because it feels like optimisation, without evidence it moves the outcome you actually want.

Where to focus in 2026

If you’re deciding where to spend limited development time on structured data, this is roughly the order of priority we’d suggest:

1. Organization schema on your homepage — name, logo, contact details, sameAs links to your social and business profiles. Low effort, clear identity signal, broadly useful.
2. Product or Service schema for anything you sell, with accurate pricing, availability and review data if you genuinely have reviews to cite. Never fabricate ratings data — this is one of the few schema misuses Google actively penalises.
3. Article schema for blog and editorial content, with correct author, publish date and modified date — useful for freshness signals across both traditional and AI search.
4. An entity graph approach (@graph with @id cross-references) if your site has any complexity — multiple locations, a parent/subsidiary structure, or a mix of content types — so the relationships between entities are explicit rather than implied.
5. Skip or deprioritise FAQPage and HowTo schema purely for the rich-result payoff — that payoff has been removed or heavily restricted. If FAQ content genuinely helps your readers, keep it, marked up or not; just don’t budget development time against a rich snippet that no longer appears.

The takeaway

Structured data hasn’t stopped mattering — it’s just stopped being a universal SEO cheat code, if it ever really was one. The platforms with clear, on-the-record guidance (Google AI Overviews, Bing Copilot) are worth building for. The rest is either genuinely unproven or, in FAQ schema’s case, a feature that quietly stopped existing while a lot of checklists didn’t notice. Spend the effort on clear entity markup and let your actual content quality do the rest of the work.

We build structured data into every site we deliver as standard, but we review what’s implemented against what’s currently confirmed to matter — not what was best practice three years ago. If you’d like us to audit your existing schema, get in touch.