Google’s guide to optimizing for generative AI features says structured data isn’t required for AI search. It also recommends continuing to use it as part of your SEO strategy because it can help with eligibility for rich results. Google didn’t say to abandon the tactic.
I think the big challenge with AI citability, or how likely AI tools are to reference and cite your content, is misreading the data or the statement. A platform announcement isn’t yet a website change request. Someone still needs to figure out what it means for your pages.
Want the full conversation? Watch Richard talk through AI search, structured data, and how to make sense of the latest SEO advice on YouTube. Or keep reading for a closer look at what deserves a change on your website.
Key Takeaways
Start with what you want AI visibility to achieve for your business.
Google’s guidance about AI requirements isn’t a blanket recommendation to remove structured data.
Match your response to the evidence: leave the implementation alone, run a limited test, or make a broader change.
A tactic that works for one website in one moment isn’t a universal recommendation.
How Should You Evaluate AI Citability Advice?
I like to simplify these decisions. Traditional search and generative search are trying to answer a user’s question and connect them to the best possible answer. That’s the context I want to keep in mind before getting into a tactic.
Use this worksheet to figure out what an announcement actually means for your website before deciding whether anything needs to change. The example follows a site considering whether to remove FAQ schema after reading Google’s guidance.
Worksheet FieldWhat to Write DownExample: Google’s Structured Data GuidanceYour goalWrite down what you want AI visibility to achieve for the business.Increase awareness of our business among people asking questions we can answer.Exact statementRecord what the platform actually said, using its original wording or an accurate summary. Keep that separate from the headline somebody wrote about it or the recommendation they attached to it.Google’s guide says structured data isn’t required for generative AI search, but recommends continuing to use it as part of an overall SEO strategy because of its role in rich-result eligibility.Affected product or featureName the specific search feature being discussed. Don’t assume a statement about one feature explains how another handles your content.Google’s AI Overviews and AI Mode. This guidance doesn’t establish how ChatGPT or other AI platforms handle structured data.Affected site elementsIdentify what the statement concerns on your pages. FAQ structured data, the FAQ answers themselves, and the headings organizing your content are different things.The decision concerns FAQ schema, not deleting the questions and answers people read or the headings that organize them.Benefit actually lostRecord the specific benefit being removed, if any.None established by this statement. Not required doesn’t mean deprecated or no longer useful.Remaining valueWrite down what the implementation still does for a person or a tool trying to understand the page. Structured data, for example, gives machines an explicit description of the relationship between questions and answers instead of leaving that relationship entirely implicit in the page content.Relevant structured data can still support eligibility for applicable rich results. That doesn’t mean every site qualifies for FAQ rich results or that FAQ markup guarantees AI citations.Evidence needed before changing anythingIdentify the outcome you want to improve and the information available to judge it.Review the affected pages and topics. Google’s Generative AI performance report in Search Console provides impression data for AI Overviews and AI Mode. V9’s GEO Grader can offer another perspective on AI search visibility. Neither establishes that a particular markup change caused an appearance.
I think of AI visibility primarily as awareness, with some potential for conversion. More visibility doesn’t automatically mean more traffic or revenue.
There’s another distinction in the data: is your visibility low because there isn’t much demand for what you offer, or because you’re missing opportunities to appear in relevant searches? Those are two different problems, and I’d approach them differently.
I look at the prompts people are asking, the topics and pages appearing in results, and what the available data suggests about interest in those topics. If other businesses are appearing for relevant questions and yours isn’t, that’s an opportunity worth investigating. But low visibility alone doesn’t tell you whether demand is the issue, especially when the available data is limited.
Should You Keep, Test, or Change Your Approach?
Once you’ve figured out what the announcement actually affects, I’d narrow the decision down to three options: leave it alone, run a limited test, or make a broader change.
Leave the Implementation Alone
For a site considering schema removal, I wouldn’t treat Google’s guidance alone as a reason to make that change.
Headings are another good example. They organize information and help readers find what they’re looking for. Their usefulness doesn’t disappear because someone says they aren’t a ranking factor.
Those headings also support content chunking, which is essentially about making your page skimmable. Readers should be able to find the answer to their question or the information that’s helpful to them in that moment.
Run a Limited Test
Sometimes there’s a specific improvement worth trying without treating it as a requirement for every page.
In our work, we’ve seen clients add FAQ sections to a page and get no results. After adding structured data to those FAQs, we started seeing inclusion in AI search results. Seeing results afterward doesn’t prove the structured data caused the change, but it’s something worth testing further.
I wouldn’t use that observation to argue that schema is required for Google’s AI features.
When someone says this is what works, I always think that’s an incomplete sentence. This is what works for you in this moment. It isn’t universal. I want to understand what might be beneficial and try it, rather than assume somebody else’s result settles the decision for your website.
Make a Broader Change
Sometimes improving your brand visibility in AI search takes more than adjusting a single page. I like to tell clients that the goal should be for your brand online to at least match, if not exceed, the authority you believe it has. But what you know about your business isn’t always reflected in what people can find about it online.
Improving your online presence can require work across your on-site content and how other people talk about your business. Who’s talking about you matters. How they’re talking about you is equally important. That means earning genuine coverage, reviews, and recommendations, not manufacturing them.
At V9, our SEO work considers the brand’s broader online presence. I want people and bots to see a business that is authoritative, knowledgeable, helpful, and worth doing business with.
How Can You Make Smarter AI Search Decisions?
Good SEO and AI search strategies aren’t about following every new recommendation. They’re about understanding what matters for your business and making thoughtful changes that support your goals. It’s how I’d approach any decision about your website.
If you’re unsure which SEO or AI search updates deserve your attention, talk to a V9 expert. We can help you evaluate what’s happening, identify where changes could make a difference, and focus on the work that supports your business goals.
FAQs About SEO Updates and AI Search
Should I remove FAQ schema if Google says it isn’t required for AI search?
Not based on that statement alone. Google’s guidance says structured data isn’t required for generative AI search, but recommends continuing to use relevant structured data as part of an overall SEO strategy. FAQ schema doesn’t guarantee AI citations, and not every website is eligible for FAQ rich results. I’d consider whether the markup still serves a useful purpose before removing it.
How can I check why an AI tool recommends my brand?
You can ask an AI tool which sources support its recommendation, but treat the response as a starting point, not a reliable explanation of how it reached its answer. Review any sources it provides and check whether reviews, third-party recommendations, or other brand mentions actually support what the tool is saying.
How much can a small site learn from limited data?
It’s harder to draw meaningful conclusions when your website doesn’t generate much traffic or visibility data. I’d focus on what you do have, look for patterns, and avoid making big decisions based on a small sample. Limited data doesn’t mean you can’t improve your strategy, but it does mean you need to be realistic about what the evidence can tell you.
