How to get your business cited in AI search: what the research shows
Search has quietly changed what winning looks like. In the first four months of 2026, 68% of Google searches in the US ended without a click, up from 60% in 2024, according to SparkToro's analysis of clickstream data. More questions are answered on the results page or inside an AI tool.
For a B2B business that means two goals, not one: rank, and be the source the answer names. This guide pulls together what the best available research says about the second.
Why does being cited in AI answers matter?
Because the gap between cited and not cited is large and growing. Seer Interactive tracked 5.47 million queries across 53 brands. On informational searches where Google showed an AI Overview, brands cited in it earned a 2.07% click-through rate, against 0.94% for brands that weren't: about 120% more clicks per impression.
For brands left out, things got worse through the year. Seer found click-through rates on those searches fell 67% over 2025. Seer is careful to say this is correlation, since well-known brands are more likely to be cited in the first place, but the direction is hard to ignore.
Do you need special "AI optimisation" to be cited?
Not for Google. Its guidance on AI features and your website says there are "no additional requirements to appear in AI Overviews or AI Mode" and that you don't need new files or special markup. The basics still decide it: pages that can be crawled, content in text rather than locked in images, good internal links, and structured data that matches what's on the page.
For ChatGPT search, OpenAI's documented requirement is that its crawler, OAI-SearchBot, can access your site. Claims that registering with Bing or using IndexNow will get you into ChatGPT are not confirmed by OpenAI, so treat them as nice-to-haves.
The practical first step is boring but important: check your robots.txt and any bot-blocking settings at your host or CDN aren't shutting AI search crawlers out.
What makes content more likely to be quoted?
The most rigorous evidence comes from a Princeton-led research team, published at KDD 2024. They tested how different edits changed a page's visibility in generative engine answers.
| Change to the content | Effect on visibility in AI answers |
|---|---|
| Adding quotations from relevant sources | About +41% |
| Adding statistics | About +31% |
| Improving fluency and readability | About +28% |
| Citing sources | About +27% |
| Using more technical terms | About +18% |
| Keyword stuffing | About -8% |
The pattern makes sense. An AI answer needs something it can lift and attribute: a number, a quote, a clearly sourced claim. Pages that read like old-school SEO copy give it nothing to use.
Does original data help?
Yes, but only if it is easy to find and use. Kevin Indig's July 2026 analysis of AI citations found that primary research made up just 2.7% of cited pages but earned 8.4% of citations, averaging 11.3 citations a page against 3.4 for everything else. One benchmark from Fivetran earned 44 citations on its own.
Indig's conclusion is that "AI does not reward 'original data' by default". Data gets cited when it answers a specific comparison, puts the headline result early, explains its method, uses a table, and lives at a URL that doesn't move. Data buried in a gated PDF does very little.
Where do AI tools find B2B information?
Mostly not on company websites. Meltwater analysed 9.5 million citations across 16 B2B categories in ChatGPT, Claude and Microsoft Copilot. Company websites made up 18.7% of citations, while third-party and user-generated platforms made up 47.5%. YouTube ranked first and LinkedIn second.
The detail that matters most for smaller firms: about 75% of LinkedIn citations came from personal profiles, not company pages. Your founders and senior people posting useful, specific content is part of your search strategy now.
A checklist for B2B teams
- Make sure AI search crawlers can reach your site. Check robots.txt and your host's bot settings.
- Answer the question in the first few lines, then explain. A short summary at the top gives AI something to quote.
- Use headings that match the questions buyers ask.
- Back claims with statistics and name your sources, linking to the original rather than to an article about it.
- Include a quote from a named expert or an official source where it adds weight.
- Use tables for comparisons. They are easy for both people and machines to read.
- Keep URLs stable, and update a page rather than publishing a new one on the same topic.
- Share the insight on LinkedIn from real people, not just the company page.
None of this replaces having something worth saying. But if you do, these steps make it much more likely that you're the business AI names. If you're wondering whether AI-written content puts you at risk while you do this, our guide on whether Google penalises AI content covers the rules. Or talk to us about putting it into practice.
Sources
- Search Engine Land: Google zero-click searches reach 68% in early 2026, 9 June 2026
- Seer Interactive: AIO impact on Google CTR, 2026 update
- Google Search Central: AI features and your website
- eSEOspace: Bing Webmaster Tools and IndexNow for AI search, what's confirmed
- Aggarwal et al.: GEO, Generative Engine Optimization, KDD 2024
- Search Engine Land: Why most original data never gets cited (Kevin Indig), 8 July 2026
- IT Brief UK: LinkedIn ranks second in AI citations for B2B searches, 1 June 2026
Alfie co-founded Agpo and leads its strategy and sales work, helping B2B teams line up their marketing with what their salespeople actually need.
How we write: we use AI tools to help research and draft. Every post is fact-checked, edited and approved by Alfie or Will before it's published, and every statistic links to its source.
Sales team asking for more than you can make?
Tell us what's missing. We'll reply within one working day with how we'd tackle it.
Submit your request