FAQ and HowTo schema after Google's deprecation: what still works?

Google has pulled FAQ and HowTo rich results out of the search results. The short answer to whether you should still use that schema: the markup stays safe and useful, but no longer for the visible block in Google. The value shifts somewhere else, namely to how well AI engines such as ChatGPT, Perplexity and Google’s AI Overviews can understand and cite your content. In this article you will read which structured data still pays off, including FAQ structured data, and which types you can safely let go.

What exactly changed for FAQ structured data

Structured data, or schema markup, is code you add to a page to tell search engines what is on that page. For years, companies used FAQ and HowTo schema mainly for one reason: it earned extra space in the search results. You got expandable questions or step-by-step instructions under your link, and that literally pushed competitors down the page.

That visible reward is gone. Google no longer shows FAQ and HowTo blocks in the regular search results. For many marketers that feels like a reason to strip the markup out immediately. That is an understandable, but mistaken reflex.

Important to understand: the markup itself is not broken. It causes no errors, it does not hurt your rankings and Google keeps reading the data. What disappeared is only the cosmetic display. So the question is not whether the schema still works, but where its value has moved to.

Why structured data is becoming more important, not less

The search world is shifting from blue links to answers. More and more people ask their question directly to an AI assistant and get a summarised answer with a handful of sources. That fundamentally changes the role of schema.

AI engines usually do not read your JSON-LD one to one as hard data. But the discipline schema enforces makes your content easier to extract. An FAQPage structure forces you to pair one question with one self-contained answer. A HowTo forces you to lay out steps clearly and in order. That structure is exactly what a language model needs to cite your content reliably instead of paraphrasing it or leaving it out.

So the win no longer sits in schema as a snippet trick, but in schema as a way to make your entities, facts and relationships unambiguous. A machine that understands without doubt who you are, what you offer and which questions you answer is a machine that includes you correctly in an answer more often. That is the core of what we, as an SEO specialist, mean by optimising for both rankings and citations.

There is something else at play. An AI engine composing an answer is managing risk: it picks sources that are easy to verify and hard to misread. The more ambiguity that surrounds your brand and content, the greater the chance a model skips you in favour of a source that is structured more clearly. Schema reduces that ambiguity. It is no guarantee of a citation, but it lowers the threshold at which an engine decides to mention you after all. In a world where most B2B competitors neglect their structured data, that is a difference you can create relatively cheaply.

Which structured data still pays off

Not every schema type has been hit equally. It helps to think of them in three groups.

Schema that still renders visually

A number of types still deliver visible enrichment in the search results and therefore deserve your time:

• Organization and LocalBusiness. These capture your brand as a recognisable entity, with name, logo, address and contact details. It is the foundation both Google and AI engines build on to recognise you as a source at all. How to link that entity explicitly to the Knowledge Graph is covered in sameAs schema and entity linking.
• Product and Offer. For concrete offerings, price, availability and features remain useful, also in a B2B context where you describe services or packages.
• Breadcrumb. Shows your site structure in the results and helps machines understand the hierarchy of your content. Read how to set that up in breadcrumb schema for your site structure.
• Article and Author. Ties content to an author and a publication date, which weighs in on how credible your content comes across.

Schema that no longer gives a snippet, but keeps its value

FAQPage and HowTo fall in here. The expandable block is gone, but the markup remains a cheap form of insurance. It takes little effort, carries no risk and reinforces the Q&A and step structure that AI engines like to pick up. For the practical approach per type, read FAQPage schema for AI citations and HowTo schema for AI.

The nuance: adding the schema only makes sense if the underlying content holds up. One question with one sharp, self-contained answer is valuable. Ten vague questions that exist only so you can place markup help nobody. Structure is the means, not the goal.

Schema you can safely let go

If you once added FAQ or HowTo markup purely to score the visible block, and the content behind it is weak or padded, you gain nothing by keeping it. Cut the hollow Q&A instead and invest that time in a genuinely good answer on the page itself. Markup around empty content is wasted effort, not insurance.

How to tackle this in practice

You do not have to overhaul everything at once. A level-headed order works best.

Start with the foundations. Make sure your Organization schema is correct across your whole site and that your most important service and product pages carry the right types. This is the layer that determines whether you are recognised as an entity at all.

Keep your existing FAQ and HowTo markup, but use the deprecation as a reason to revise the content behind it. Write every answer so that it still holds up without the question, in the language your buyers actually use. That makes it extractable for AI and readable for humans.

Validate what you have. Faulty markup delivers no advantage and can muddy your signals. How to spot and fix common errors is covered in validating schema markup and fixing errors.

Then work in layers. Once the foundations are in place, pick the schema per page type that matches the function of that page. A service page calls for different markup than a knowledge article or a product overview. Do not try to stuff every page with as many types as possible; that makes your signals messier instead of stronger. One well-chosen, correctly filled type per page does more than five half-filled ones. Plan this clean-up as a recurring check, because your site grows and markup goes stale over time without anyone noticing.

And remember where the real lever sits. Schema is a readability layer on top of good content, not a replacement for it. No type of structured data saves a page that gives no clear answer. The order is always: first content that truly answers a question, then the markup that helps machines see it.

Schema in the wider growth engine

It is tempting to treat structured data as a technical checkbox. But within a well-considered approach it is a link in something bigger. For us, SEO is the acquisition layer of one orchestrated growth engine, optimised for pipeline and not for vanity ranking numbers. Schema that makes your brand readable for AI engines is only valuable if that visibility also leads to your service pages and ultimately to concrete enquiries.

That is why we never look at one type of markup in isolation, but at how your entire site helps machines understand and cite you correctly. Anyone who wants to go beyond schema alone will find the foundation in our guide on what SEO is and how it works.

Ready to make your structured data future-proof?

FAQ and HowTo schema are not dead, they have simply changed roles. The companies winning now treat schema as a readability layer for AI and not as a snippet trick. Want to know which structured data genuinely moves your B2B site forward and how to get found and cited by AI search engines? Get in touch and we will look together at where your levers are.