Microsoft Copilot brand tracking: see how Copilot mentions you
Daily Microsoft Copilot answers, with the brands named and the sites cited.
- What is the best expense management software for a 50-person company?
- Which cybersecurity vendors are recommended for small businesses?
- Compare [your brand] with [competitor] for Microsoft 365 users.
- What are good alternatives to [competitor] that integrate with Teams?
- Summarise reviews of [your brand].
How Microsoft Copilot builds its answers
- 01Copilot answers web questions using Bing-based retrieval. Microsoft's Bing Webmaster Tools AI Performance report, released in public preview in February 2026 according to trade coverage, reports citations from Copilot, AI summaries in Bing and some partner integrations.
- 02That report shows total citations, cited pages and grounding queries, the phrasings Bing uses to retrieve pages, which are not necessarily what the user typed. It does not show clicks, and it cannot be filtered to Copilot alone.
- 03Because retrieval runs through Bing, being indexed in Bing matters for Copilot visibility in a way that Google indexing does not cover.
- 04Copilot appears across Windows, Edge, Microsoft 365 and the web, so the audience skews toward workplace and desktop users.
- 05Answers show source links, so the citation mix (publishers, review sites, forums) is directly observable.
What GeoGen measures in Microsoft Copilot
- Visibility
- The share of Copilot answers to your prompts that mention your brand, tracked daily.
- Average position
- Where in the answer your brand appears when it is named.
- Citations
- The domains and URLs Copilot links to, which show which sources Bing-based retrieval favours in your category.
- Sentiment
- A 0-100 score for how Copilot describes your brand.
- Competitors
- Auto-detected competitors named next to you, per prompt.
Copilot brand tracking records what Microsoft Copilot says about your company when someone asks a buying question, which competitors it names, and which sites it cites. GeoGen reads Copilot as a user sees it and repeats the run every day.
How does Copilot find sources?
Copilot’s web answers lean on Bing. Microsoft’s Bing Webmaster Tools now offers an AI Performance report (public preview, introduced in February 2026 per trade coverage) that lists how often your pages are cited across Copilot, Bing AI summaries and some partner integrations. It reports grounding queries, the retrieval phrasings Bing generated, which are not always what the user typed. It does not report clicks and does not isolate Copilot.
Microsoft’s own pages on how Copilot ranks and cites were not available to verify in detail, so treat the finer mechanics as unconfirmed and measure outcomes instead.
What can a brand influence?
- Bing indexing. Verify your site in Bing Webmaster Tools, submit a sitemap and confirm Bingbot is allowed.
- Citable pages. Clear, factual pages with pricing, specifications and comparisons give retrieval something to quote.
- Third-party presence. As on other engines, review sites and publishers that Copilot cites are often where recommendations come from.
- Workplace framing. Integration and compliance content (Teams, Microsoft 365, SSO) matches the questions Copilot users ask.
What GeoGen measures for Copilot
Prompts run daily per language and country. GeoGen records each answer and reports visibility, mentions, average position, sentiment, citations and detected competitors. The recommended actions turn citation gaps into tasks, such as getting listed on a source Copilot repeatedly cites. LLM traffic tracking can then show whether AI referral visits and crawler visits to your site follow.
Prompts worth tracking
Shortlist and comparison prompts for B2B categories, “alternatives to” prompts and integration-specific questions. Keep a handful of branded prompts to monitor how Copilot summarises you.
Combining Bing’s report with prompt tracking
The two data sets answer different questions. Bing’s AI Performance report starts from your pages and tells you how often they were cited, and the grounding queries behind those citations. It is a good source of phrasing ideas, since it shows how Bing translates user questions into retrieval searches.
Prompt tracking starts from the question and tells you what Copilot actually said: whether you were named, in what position, with what tone, and which competitors shared the answer. It also covers prompts where you were not cited at all, which a citation report cannot show.
A practical routine is to export the grounding queries monthly, turn the most relevant into tracked prompts, and watch whether visibility moves after you publish or update the matching pages. Keep each change small and dated, so you can link an effect to its cause.
Next steps
Copilot is one of 10 engines GeoGen reads as a user sees them; the number you can track at once depends on the plan. For tactics, read the Copilot AI search optimization guide, then compare with ChatGPT and Gemini, or view the monitoring overview.
Frequently asked questions
Does Bing Webmaster Tools replace Copilot tracking?
It complements it. The AI Performance report shows how often your pages are cited and the grounding queries behind them, but it lacks clicks, cannot isolate Copilot and does not show which competitors were named. A tracker records the answers themselves.
Do I need to be indexed in Bing to appear in Copilot?
Copilot retrieval for web answers is built on Bing, so indexing in Bing is the sensible baseline. Submit your sitemap in Bing Webmaster Tools and check that Bingbot is not blocked.
Is Copilot tracking relevant for B2B brands?
Often more than for consumer brands, because Copilot sits inside Windows, Edge and Microsoft 365, where many workplace research tasks start. Test with the B2B shortlist prompts your buyers actually use.
What is a grounding query?
A grounding query is the search phrasing Bing uses to retrieve pages for an AI answer. It may differ from what the user typed, which is why prompt-level tracking and retrieval reports tell different halves of the story.
How does GeoGen read Copilot?
As a user sees it, the public answer, rather than via a model API. One run of one prompt costs 1 credit.
Other engines
Be the brand AI recommends.
Start tracking how ChatGPT, Google AI Overviews, Gemini, Perplexity, Claude and 11 more AI engines talk about you. Results in minutes.