Glossary

What are grounding queries in AI search?

Grounding queries are the search questions an AI assistant uses behind the scenes to find sources before it writes an answer. In plain English, they’re the back

By Published 4 min read

Grounding queries are the search questions an AI assistant uses behind the scenes to find sources before it writes an answer. In plain English, they’re the background searches that decide whether your business is considered as evidence. If your pages appear for those background searches, you’re more likely to be cited or mentioned in AI answers. If you don’t, you can be invisible even when you rank for normal search terms.

For a business, this matters because AI assistants don’t always answer from memory. They often go out and look for supporting sources first. Those hidden lookups shape which brands, pages and facts get pulled into the final answer.

Think of it this way. A person might ask an AI assistant, “What is the best payroll software for a small business in Ireland?” The assistant may then run several background searches such as “best payroll software Ireland”, “payroll software pricing”, “payroll software reviews” or “payroll software for SMEs”. Those are the grounding queries. Your visibility depends on whether your site shows up as a useful source for those searches.

This is why grounding queries matter for getting found in AI search. You’re not only trying to rank for the exact words a customer types. You’re also trying to be present for the supporting searches the assistant uses to check facts, compare options and build confidence in its answer.

There is now a first-party source for some of this data. Bing Webmaster Tools added its AI Performance report in public preview on 1 February 2026, with query-to-page mapping added on 23 March 2026, according to Microsoft. That report includes grounding queries, which lets you see the questions where Bing and Copilot cited pages from your site. On 13 October 2026, Microsoft also added Topics Portfolio in beta, which shows the grounding queries where competitors are cited and you are not. For businesses, that’s useful because it turns a vague visibility problem into a list of missed opportunities.

In practical terms, grounding queries help you answer three business questions. First, what topics is your brand already trusted for. Second, where are competitors winning citations. Third, which pages should you improve first.

Here is a simple example. Say you run an accountancy firm in Cork. You want to be recommended when someone asks an AI assistant about “tax advice for contractors in Ireland”. The assistant may use background searches such as “Irish contractor tax deadlines”, “limited company vs sole trader Ireland” and “accountant for contractors Cork”. If your site has clear pages that answer those questions directly, with visible question-and-answer text, real bylines and honest dates, you have a better chance of being used as a source. If your site only has a generic services page, you’re less likely to be included.

This does not mean you should rewrite your copy “for AI”. Google’s generative AI optimisation guidance says this is unnecessary, and Google says its systems can understand multi-topic pages without special chunking. What matters more is whether your content answers real questions clearly, whether your pages can be accessed by AI readers, and whether your business is consistently present across the web.

It also helps to understand what grounding queries are not. They are not an “AI ranking position”. There is no single ranking slot in AI answers. They are also not the same as ordinary keyword reports in search engine optimisation. A normal keyword report tells you where you rank in search results. Grounding queries show the supporting searches an AI system used when deciding what to cite.

If you want to act on this, start by reviewing the grounding queries you can see in Bing Webmaster Tools. Then map them to the pages on your site that were cited, or should have been cited. Look for gaps in your coverage. Add or improve pages that answer those questions directly. Keep your basics in place too, such as a sitemap, your site’s table of contents, and robots.txt, your AI welcome sign, so the right AI readers can access your pages. If you want a broader view of what tends to help, see What works in AEO.

AI citations, AI visibility, Bing Webmaster Tools AI Performance report, answer engine optimisation. See the full glossary at /glossary/.

FAQs

Are grounding queries the same as the prompts people type into AI assistants?

No. The prompt is the user’s question. Grounding queries are the extra background searches the assistant may run to find sources and verify claims before answering.

Can I see grounding queries for my business?

Sometimes. Microsoft’s Bing Webmaster Tools AI Performance report, first released in public preview on 1 February 2026, includes grounding queries for Bing and Copilot citations. Query-to-page mapping was added on 23 March 2026, according to Microsoft. There is no equivalent first-party query report from Google for AI Overviews at the time of writing.

How do I improve my visibility for grounding queries?

Create pages that answer the real supporting questions around your products or services. Use plain language. Show expertise with real bylines and honest dates. Make sure AI readers can access the page. Then track which queries already lead to citations, and where competitors are being cited instead of you.

Next step

Find out where you stand in AI answers

The AI Visibility Audit tests your buyers’ real questions across five AI assistants, shows who gets cited instead of you, and ends with a phased 90-day plan.