Can I see which queries Copilot used when it cited my site?

Microsoft added an AI Performance report to Bing Webmaster Tools as a public preview on 10 February 2026 and expanded it on 16 June 2026. It reports citation activity across Microsoft Copilot, AI-generated summaries in Bing, and what Microsoft calls select partner integrations.

The part you're asking about is the grounding queries view. Microsoft defines grounding queries as "the key phrases the AI used when retrieving content that was referenced in AI-generated answers." In plain terms: when Copilot cited your page, this is roughly what it searched for to find it.

To see it, you need a verified site in Bing Webmaster Tools. If you haven't done that yet, start with setting up Bing Webmaster Tools for AI reporting. The report sits in the web interface. As of the evidence we hold (August 2026), there's no API for it, though Microsoft's Fabrice Canel has indicated API access is planned during 2026. Check the current state before you build a workflow on it.

The report exposes several views:

• Total Citations: how often your content was cited.
• Average Cited Pages and page-level citation activity.
• Grounding queries: the retrieval phrases behind those citations.
• Intents: categories such as Informational, Commercial, Navigational, Research and Local.
• Topics, Citation Share and a Compare view.

Two limits matter. First, Microsoft attaches its sampling note to the grounding queries metric specifically: the data shown "represents a sample of overall citation activity." The other metrics carry different caveats. So don't tell your team the whole report is sampled, and don't treat the query list as complete. Our page on whether Bing's AI Performance data is sampled covers this in detail.

Second, it only sees Microsoft's surfaces. It's not a ChatGPT report, and nothing in our evidence shows it covers ChatGPT. In Otterly's 2026 experiment, sites that ranked in Bing still drew effectively zero ChatGPT citations.

Most AI visibility tools, including prompt trackers, test questions the vendor or customer chose. That's a proxy. The engines don't search with the user's prompt.

• OpenAI says ChatGPT search "typically rewrites your query into one or more targeted queries" before sending them to search providers.
• Google's Gemini API documentation says grounding "automatically generates one or multiple search queries" and runs them.

So the question a buyer types and the query that retrieves your page can differ. Bing's grounding queries are the only first-party view of that hidden input anywhere in the industry. That's why they're worth the manual effort of pulling them out.

The gap between samples and first-party counts can be large. In Otterly's experiment, Microsoft's report credited one test site with 42,817 citations, while Otterly's own prompt-based tracker logged 3,934. The time windows may not match, so it isn't an exact ratio. It does show that a prompt sample sees less than the engine's own count.

Citation Share looks like a competitive metric. Microsoft says it isn't. It describes Citation Share as "an observational metric" and "not a ranking system or a competitive scoreboard." Microsoft also states that it does not expose competitor domains, represent traffic share, or assign quality scores to content.

What that means in practice:

• It's a share, not a count. The denominator is dominated by whoever already wins a query. Your share can fall while your absolute citations rise.
• It won't name competitors. If you want to know who else gets cited, see whether the report can show competitor citations.
• It isn't traffic. A citation doesn't mean a click.

Report Citation Share to stakeholders with Microsoft's framing attached. Lead with Total Citations and the grounding queries themselves, which are the more useful signals.

The report is only useful if it changes what you check or fix. A practical sequence:

1. Confirm you're eligible. Copilot grounds answers in Bing's index, as covered in how Copilot grounds answers in Bing. An empty report can simply mean Bing hasn't indexed the pages, or that Bingbot can't reach them past your CDN or firewall.
2. Export the grounding queries regularly. With no API, someone has to pull them by hand. Our guide to exporting and analyzing Bing grounding queries walks through it.
3. Compare them with the questions you test. If the real queries look nothing like your tracked prompts, your tracking is measuring the wrong input.
4. Read the cited pages. Check that the facts on them (prices, product names, definitions) are correct and present in the HTML before JavaScript runs.
5. Check other engines separately. Bing data says nothing about ChatGPT or Google. The free checker on this site gives a snapshot of whether AI engines name you for up to three buyer questions.

What isn't in your control: which queries Copilot generates, and whether it cites you for them. Microsoft doesn't publish how it ranks grounding results.

Is there an API for Bing's AI Performance report? Not as of August 2026. The public Bing Webmaster API reference shows no endpoints for citations or grounding queries, and a Microsoft Q&A thread from February 2026 confirmed there was no API. Microsoft's Fabrice Canel has indicated API access during 2026, so check before building anything. More detail is on our page about [the Bing AI Performance API](/questions/bing-ai-performance-api).

Does the report include ChatGPT citations? Microsoft says the report covers Microsoft Copilot, AI-generated summaries in Bing and select partner integrations. It doesn't name ChatGPT, and our evidence doesn't show ChatGPT data in it. OpenAI says ChatGPT search sometimes partners with other providers, but in Otterly's 2026 test, ranking in Bing produced effectively zero ChatGPT citations. See [whether the report includes ChatGPT](/questions/bing-ai-report-chatgpt-citations).

Why is my AI Performance report empty or nearly empty? The most common reasons are that Bing hasn't indexed the pages, Bingbot is blocked at robots.txt or by a firewall, or Copilot simply hasn't cited you yet. Start with index status in Bing Webmaster Tools, then test whether crawlers can reach your pages. Keep in mind the grounding queries view is a sample, so a short list doesn't mean zero activity.

Can I use grounding queries to predict what Copilot will cite next? Not reliably. Grounding queries show what Copilot searched for when it did cite you, in a sample. They don't reveal how Microsoft ranks results or which pages it passed over. Use them to learn the language buyers' questions turn into, and to check that your pages state the relevant facts clearly. Treat them as evidence of past behavior, not a forecast.

Does Citation Share tell me my rank in Copilot? No. Microsoft explicitly says Citation Share is not a ranking system or a competitive scoreboard. It's a share whose denominator is dominated by the sources that already win a query, so it can drop while your own citations grow. No engine publishes position data for AI answers, so any tool claiming a Copilot rank has nothing official to check it against.