Summary
The best way to get cited by AI is to publish original data that only you can offer, then make it easy to find and easy to quote. Put your headline figure near the top, explain what it measures, show your method and write key numbers in plain text. Don’t count on schema to do the heavy lifting. And because most AI citations come from earned media, get your findings covered by others too. Do this consistently and your AI visibility grows with your reputation as a trusted source.
How AI Engines Choose Sources
According to Semrush, AI tools get their information in four main ways: from the data they were trained on, from live web searches, from licensing deals with content partners such as Reddit, and from files or apps that users connect themselves. When it comes to recent statistics, most of the work is done by live web searches. Google’s AI Overviews and AI Mode pull from Google’s own search index, while ChatGPT combines results from OpenAI’s own crawler with other search indexes.
So if crawlers can’t reach your page, your data can’t be cited. And even once a page is found, the AI answer engine still has to choose the clearest, most trustworthy passage from many options. Winning that choice is what answer engine optimisation and generative engine optimisation are really about.
Why Original Data Gets Cited More
Numbers give AI something specific to point to
Generic advice reads the same on every site, so AI has little reason to credit any one of them. A precise figure with a date and a source is different, because it can be checked and traced back to a single origin.
This is also where many pages fall short. Plenty of articles simply repeat the same widely shared statistics, often without checking where they came from. A page that adds genuinely new figures from your own work gives readers and AI a reason to choose it and a better chance of earning AI citations.
Being first isn’t the same as being cited
There is a catch. Publishing a number first won’t always earn you the credit, because AI tends to quote whichever page explains it most clearly, even if that page didn’t create it. That’s why presenting your data well matters.
How to Present Statistics So AI Can Quote Them
Good AI content optimisation for data comes down to a few simple practices.
| Element | What to do | Why it helps |
|---|---|---|
| Headline figure | Put your strongest number near the top of the page | Readers and AI tools both see it straight away |
| Definition | Explain in one line what the number measures and who it applies to | A number with no context is harder to quote with confidence |
| Method | Add a short box with sample size, time period and collection method | It shows the figure can be trusted |
| Other findings | List them in order of strength, with strongest first | Your key points don’t get lost at the bottom |
| Visible text | Write key figures in the page text, not only in a chart | AI tools read what is visible on the page |
Above all, don’t save your best finding for a big reveal at the end. Semrush also recommends making every section self-contained, so it gives a complete answer to its heading without relying on the rest of the page. Our guide on how to structure content so AI engines can extract and cite it covers page layout in more detail.
What about schema markup for AI?
You’ll often hear that schema markup for AI is the fastest route to citations. The data, however, doesn’t support that claim. Ahrefs tracked 1,885 pages that added schema between August 2025 and March 2026 and compared them with 4,000 similar pages that didn’t. Citations in ChatGPT and AI Mode barely moved, and AI Overview citations dipped slightly.
Ahrefs also points to a separate test which found AI tools read only visible content when fetching a page live. Structured data for AI is still worth keeping, since it helps your pages appear as rich results in Google. However, it can’t make up for important figures that readers and AI tools struggle to find.
How to Get Your Data Cited by AI Beyond Your Own Website
- Use data your business already has. You don’t need a research team to get started. Product usage, pricing trends, survey responses and client results can all be turned into useful statistics. These numbers come from your own business, so they make your data-driven content hard for anyone else to copy.
- Get others to cover it. AI leans heavily on earned media, so send your findings to journalists and the trade publications your buyers read. Showing up in those outlets is good for B2B brand building too.
- Keep your story consistent. Ensure that what other websites say about your business matches the information on your own site. If your details match across the web, AI is far less likely to get confused about who you are.
- Make sure AI can reach your pages. It sounds obvious, but if your robots.txt file blocks AI or search crawlers, your data won’t get seen at all. Then log in to Google Search Console and Bing Webmaster Tools, and make sure the pages holding your data are listed there as indexed.
- Keep your figures fresh. Semrush notes that AI systems prefer citing fresh content. Whenever you have newer figures, update the page and add a clear ‘last updated’ date so readers and AI can see the information is current.
- Publish fewer, stronger studies. One solid report does more than a stack of thin posts. Our take on the content saturation paradox explains why, and our content marketing services are built around that idea.
Finally, check whether it’s working, platform by platform. Muck Rack found ChatGPT cites sources in 96% of responses but averages five links, while Claude cites in 55% and averages 13 when it does. Our guide to tracking your brand’s AI visibility shows how to measure this, and our look at whether AI search will replace SEO explains how rankings and AI answers connect. Measuring and adjusting is the heart of good AI search optimisation.
