Google changed the rules and barely told anyone. Since May 2024, AI Overviews have been appearing at the top of search results for an increasing number of queries. That box of AI-generated text above all the organic results, pulling information from various websites and presenting it as a direct answer. For some searches, the user never needs to click through to your site at all.
If you run a B2B business and your website generates leads through organic search, this matters. A lot.
The early data from studies by Seer Interactive and others suggests that AI Overviews reduce click-through rates on affected queries by 15% to 25%. For informational searches like "what is" and "how to" queries, the reduction is steeper. Google is answering the question before the user reaches your content.
But here's what most of the panic articles miss: Google's AI Overviews need sources. They need websites to pull information from. And the sites that get cited in those overviews are gaining visibility, not losing it. The question isn't whether AI Overviews are good or bad. The question is whether your content is the kind that gets quoted or the kind that gets replaced.
There are now functionally two types of B2B content. Content that answers a question, and content that is the answer.
Content that answers a question is what most B2B websites have: service pages that describe capabilities, blog posts that explain concepts, FAQ pages that address common queries. This content has been the foundation of B2B SEO for a decade. And AI Overviews are gradually making it less valuable, because Google can now synthesise those answers itself.
Content that is the answer is different. It contains original data, unique perspectives, first-hand experience, or specific expertise that can't be generated by summarising existing web pages. When Google's AI needs to cite a source for a technical specification, an industry benchmark, a case study with real numbers, or an expert opinion, it looks for content that is genuinely authoritative.
An accountancy practice broker we work with publishes content about business valuations that includes specific market multiples, sector-by-sector data, and real examples (anonymised, obviously). That kind of content gets cited because it can't be recreated from generic sources. It comes from doing the work.
Schema markup has been important for SEO for years, but it's become critical for AI Overviews. Schema is structured data that tells search engines exactly what your content is, who wrote it, what questions it answers, and how it relates to other topics.
For B2B websites, three types of schema matter most right now.
FAQ schema tells Google that your page contains questions and answers. When someone searches a question that matches your FAQ, there's a higher probability of your answer appearing in the AI Overview. The key is that the questions need to be genuine queries your audience actually searches for, not marketing-speak disguised as questions.
Article and author schema establishes expertise and authority. Google's AI prioritises content from identifiable experts at credible organisations. If your blog posts have proper author markup linking to a real person with a real LinkedIn profile and a real track record in the industry, they're more likely to be cited than anonymous corporate blog posts.
HowTo and professional service schema help Google understand what you actually do and where you do it. For B2B service businesses, this structured data connects your content to the right searches at the right time.
Most B2B websites have none of this implemented. Their content management system might generate basic page-level schema, but the detailed, content-specific markup that AI Overviews rely on is missing entirely.
The way you structure your content directly affects whether it gets pulled into AI Overviews. Google's AI looks for content that follows patterns it can parse and attribute.
Conversational Q&A works well. Not the forced "FAQ section" bolted onto the bottom of a service page, but genuine questions woven into the content with clear, direct answers. When you write "The average cost of a B2B data build for a list of 5,000 verified contacts is between £2,000 and £5,000, depending on the level of enrichment and verification required," you're giving the AI exactly the kind of citable statement it looks for.
Specific, attributable claims perform better than vague assertions. "48% email click rates across three sector-specific campaigns" is citable. "Significantly improved email performance" is not. Numbers, timeframes, and measurable outcomes give AI something concrete to reference.
Clear entity relationships help too. When your content explicitly connects the author (a named expert), the organisation (your company), the topic (a specific marketing challenge), and the evidence (real results), the AI can build a confident citation.
There's a new acronym forming in the SEO world: GEO, or Generative Engine Optimisation. It's distinct from traditional SEO because it focuses on how AI systems select and cite sources rather than how search engines rank pages.
GEO principles are still emerging, but the core ideas are clear. Authority matters more than keyword density. Original research and data outperform summarised information. Structured content with clear hierarchies is preferred over long, unstructured prose. Expert attribution with verifiable credentials beats anonymous publishing.
For B2B businesses, this is actually good news. The companies with genuine expertise, real client results, and specific industry knowledge have a natural advantage over content farms and generic AI-generated articles. Your lived experience is exactly what these AI systems need to cite.
If your website generates any meaningful portion of leads through organic search, here's what I'd prioritise.
Audit your highest-traffic pages. Check which ones are answering questions that AI Overviews now handle. If your traffic on those pages has dropped in the past six months, you've already been affected.
Add proper schema markup. At minimum, FAQ schema on your top content pages, article schema on every blog post, and organisation schema on your about and service pages. This isn't a weekend project for most businesses, but it's the kind of infrastructure investment that pays off for years.
Rewrite informational content to include original data. Take your generic "what is" posts and add your own numbers, case studies, and expert perspective. Make them citable, not just readable.
Build a FAQ content strategy based on what your actual prospects ask, not what keyword tools suggest. The questions real buyers ask are the questions AI systems will try to answer.
We've been building this kind of content infrastructure for clients as part of our strategy work. It sits at the intersection of content creation and technical implementation, and it's becoming more important with every Google update.
The businesses that adapt their content for this new reality will maintain their search visibility. The ones that keep publishing the same generic blog posts will gradually disappear from the results that matter.
Martin Dugan, AA2
