ChatGPT now processes over 10 million search queries per day. Perplexity handles another 4 million. Google's AI Overviews appear on roughly 40% of informational searches. Microsoft Copilot is integrated into Edge and Bing. Gemini is embedded in Android and Google Workspace.
The way people find information has changed, and your content needs to change with it.
When someone asks ChatGPT "who are the best B2B marketing agencies in Wiltshire," the answer doesn't come from a search results page. It comes from the AI synthesising information it's gathered from across the web. The websites it cites, the businesses it mentions, the facts it presents, are all determined by how well that content was structured for AI consumption.
If your website was built for Google's traditional search algorithm and you haven't touched it since, you're invisible to a growing share of your potential audience.
Large language models don't read websites the way humans do. They don't browse. They don't click around. They process structured information and make decisions about relevance, authority, and accuracy based on patterns in the data.
When an LLM encounters your content, it's looking for several things. Clear, direct answers to specific questions. Identifiable expertise (who wrote this, what's their authority on the subject). Structured data that confirms the semantic meaning of the content. Consistent entity relationships (this person works at this company, which provides these services, in this location).
The content that performs well in AI-powered search is content that makes these signals explicit rather than implicit. A blog post that mentions the author's name, their role, and their company in the first paragraph gives the LLM more to work with than an anonymous post on a corporate blog. A page with FAQ schema that directly answers common questions provides structured signals that raw prose doesn't.
This isn't about tricking AI systems. It's about making your content easier for them to understand and accurately represent.
I wrote about schema markup back in February in the context of Google's AI Overviews. The same principles apply to every AI-powered search tool, but the stakes are higher.
Schema markup is structured data that sits in your page's HTML and tells AI systems exactly what your content is about. Think of it as a machine-readable summary of your page's meaning.
For a B2B service business, the essential schema types are:
Organisation schema that defines who you are, where you're based, what services you offer, and how to contact you. This is the foundation. Every AI system that mentions your business draws on this data.
Person schema for your team's key people. Author information, credentials, areas of expertise. LLMs weight content more heavily when they can attribute it to a verifiable expert.
FAQ schema for every page that answers common questions. This is the single most impactful change most B2B websites can make for AI visibility. When Perplexity is answering "how much does B2B data enrichment cost," it's looking for pages with clear question-and-answer structures. FAQ schema makes those structures explicit.
Service schema that describes what you offer, the areas you serve, and the outcomes you deliver. This connects your business to the queries people ask AI about.
Article schema on every blog post, including author, publication date, topic, and word count. This helps AI systems assess recency and relevance.
Most B2B websites have either no schema or only basic organisation schema that their CMS generated automatically. The difference between that minimal markup and a comprehensive schema implementation is the difference between being occasionally mentioned by AI and being consistently cited.
AI search tools process natural language queries. "What's the average cost of a B2B email campaign?" is a typical question someone might ask ChatGPT. If your website answers that question directly, in natural language, with a clear and specific response, you have a chance of being cited.
The key word is directly. AI systems don't want to parse a 2,000-word blog post to extract a buried answer from paragraph seventeen. They want content that poses the question and provides a clear answer within the immediately following text.
This doesn't mean turning your website into a list of FAQs. It means structuring your content so that natural questions are addressed explicitly. Instead of writing "our email campaigns are competitively priced and tailored to each client's needs," write "a typical B2B email campaign for a list of 2,000 to 5,000 verified contacts costs between £1,500 and £4,000, depending on the level of segmentation, personalisation, and sequence length."
The first version is marketing language. The second version is an answer. AI systems cite answers.
LLMs build understanding through entity relationships: connections between people, organisations, places, topics, and services. The stronger and more consistent those relationships are across your web presence, the more confidently AI systems can represent your business.
Consistency matters. If your website says "AA2 is based in Wiltshire" and your LinkedIn says "Swindon, UK" and your Google Business Profile says "Royal Wootton Bassett," the AI has three slightly different data points for the same thing. Align them. Use the same descriptions, the same terminology, and the same facts everywhere your business appears online.
Internal linking reinforces entity relationships. When your blog post about email marketing links to your email campaign service page, and that service page links to a case study, and the case study links to your about page with author credentials, you're creating a web of connected information that AI systems can follow and verify.
External references strengthen authority. When industry publications, directories, or partners mention your business in context, those citations reinforce the AI's understanding of who you are and what you do. This is the AI-era equivalent of backlinks, but the mechanism is different. It's not about PageRank. It's about entity confirmation.
If you want your content to perform well in AI-powered search, here's what to prioritise over the next quarter.
Audit your schema markup. Install a schema validator browser extension and check your key pages. If you see only basic organisation schema (or nothing at all), you have an immediate opportunity.
Add FAQ schema to your top ten pages. Identify the questions each page answers, write clear Q&A pairs, and implement the structured data. This can be done in a few days and the impact is often visible within weeks.
Rewrite your service descriptions to be specific and factual. Replace vague marketing language with concrete details: what you do, who you do it for, what it costs, what results you've achieved. AI systems cite specifics, not generalities.
Ensure entity consistency. Check that your business name, location, contact details, and service descriptions are identical across your website, Google Business Profile, LinkedIn, industry directories, and any other web presence.
Create content that directly answers questions your market asks. Not "pillar content" designed to rank for broad keywords, but specific, detailed answers to specific, practical questions. "How do I warm up a cold email domain?" is a better content target than "cold email best practices" because it matches how people actually query AI systems.
We've been building this kind of content infrastructure for clients as part of our strategy work. It sits at the intersection of SEO, content creation, and technical implementation. The businesses that invest in it now will be the ones AI systems recommend in 2026.
Martin Dugan, AA2
