Home / Learn / Get Cited by Copilot
Learn

How to get cited by Microsoft Copilot

Microsoft Copilot has no search index of its own. It answers from Bing's index, retrieved by Bingbot, which is why Microsoft folded Copilot citation tracking straight into Bing Webmaster Tools instead of shipping a separate Copilot dashboard. That tool went from nothing to a real reporting suite in five months: an AI Performance report in public preview from 10 February 2026, then four more views, Intents, Topics, Citation Share and Compare, added 16 June 2026. Optimizing for Copilot is not a new discipline. It is Bing SEO, measured for the first time. And the complication, visible in a competitor's own citation data, is that Copilot's answers skew hard toward a handful of very large commercial sites.

One crawler, one index, one report

Microsoft’s own description of the February 2026 preview names what it covers directly: it tracks citations across "Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations." There is no separate Copilot crawl and no separate Copilot ranking system, because there isn’t one. The retrieval layer is Bingbot, the same crawler that has indexed the web for classic Bing search for over a decade, and the same access rules apply, covered in the guide to AI crawlers. Microsoft states plainly that "Bing respects all content owner preferences expressed through robots.txt and other supported control mechanisms," which means the crawl-access audit that decides whether you rank in Bing search is the same audit that decides whether Copilot can cite you at all. Say that to a client who arrives asking for a "Copilot strategy": there isn’t one, separate from a Bing strategy, and treating it as one is how budget gets spent on the wrong layer.

What the report actually shows, and what it deliberately doesn’t

Total Citations, per Microsoft, "shows the total number of citations that are displayed as sources in AI-generated answers during the selected time frame." Average Cited Pages shows "the average number of unique pages from your site that are displayed as sources in AI-generated answers per day." Grounding Queries lists "the key phrases the AI used when retrieving content that was referenced in AI-generated answers," but Microsoft caveats that number immediately: "the data shown represents a sample of overall citation activity." Take that caveat at face value. It is not a complete query log, and a content plan built on a grounding-query list as if it were exhaustive is a plan built on a fraction of the real traffic pattern. Page-level activity and a visibility trend line round out the report, enough to see direction, not enough to reconstruct every branch of demand, which is roughly the same limit Google’s own generative AI performance report carries.

June 2026 made the competition visible

Four features shipped 16 June 2026. Intents sorts grounding queries into categories, which Microsoft names as "Informational, Commercial, Navigational, Learn and Solve, Research, Creation, Local, and more." Topics groups related queries into themes: "instead of analyzing visibility one query at a time, publishers can now start understanding which larger subject areas are driving citation activity." Then Citation Share, defined by Microsoft as "the percentage of citations attributed to your site out of all citations shown across all sites for that same grounding query." Sit with that one, because no other engine publishes a native share-of-voice number per query. Google’s report hands you your own impressions with nothing to measure them against, and OpenAI publishes nothing at all for ChatGPT beyond what third-party tools infer, a gap covered in how to get cited by ChatGPT. Bing telling you your percentage against every other cited site, for the exact query that triggered it, is a genuinely different kind of data, even labelled a preview that is still evolving. Add Compare, which overlays a prior period on the current one, and a team can finally ask whether a page rewritten in March moved the number, instead of eyeballing two dashboard screenshots side by side.

Who is actually winning that percentage

Ahrefs pulled the answer at scale. Its Brand Radar analysis of more than 3 million US queries, updated 21 July 2026, ranks the fifty most-cited domains in Copilot: Amazon leads at 14.4% of all citations, Walmart holds 10.0%, Wikipedia sits at 9.5%. All three carry Domain Ratings in the 90s, 96, 92 and 97, and the study says plainly why: "with Microsoft now pushing shopping and checkout straight into Copilot chats, the commercial slant of its sources matters more than ever." A meaningful share of Copilot’s citation inventory is structurally reserved for a small number of retail and reference giants on broad commercial queries. No amount of on-page work moves that ceiling for a smaller brand competing head-on with Amazon on the same term.

That is the honest limit, and it is also where Intents and Topics stop being a novelty and start being a targeting tool. A query Bing files as Commercial and routes toward a wide product category is not winnable against a Domain Rating in the 90s. A query filed as Local, Learn and Solve or Research, on a specific subtopic Amazon has no page for, is winnable, and it is now possible to check that split against your own grounding queries before spending the budget instead of after.

What actually moves it

A 30-day plan

Week one, verify Bingbot access and a clean robots.txt, then confirm Bing Webmaster Tools has the AI Performance preview enabled for the property. Week two, pull the Intents and Topics breakdown for your own grounding queries and separate what is Commercial at a scale you cannot win from what falls into Local, Research or Learn and Solve, where you can. Week three, restructure and submit content for the winnable buckets through IndexNow. Week four, run Compare against the pre-project baseline and read Citation Share on the specific queries you targeted, not the site-wide average, which will hide the movement that actually happened.

Sources

The AI Performance launch details come from Microsoft’s own post, Introducing AI Performance in Bing Webmaster Tools, Public Preview, 10 February 2026. The four new views are documented in New AI Visibility Insights in Bing Webmaster Tools, 16 June 2026. The citation-share data by domain comes from Ahrefs, The 50 Most-Cited Websites in Copilot, updated 21 July 2026. IndexNow is documented directly at bing.com/indexnow.

Common questions

Is optimizing for Copilot different from optimizing for Bing search?

No. Microsoft’s own description of the AI Performance report names Copilot, Bing’s AI summaries and select partner integrations as what it tracks together, the retrieval layer is Bingbot in every case, and the access rules are the ones that already govern classic Bing SEO.

Can I see the exact query that got my page cited in Copilot?

Not completely. Grounding Queries lists the key phrases involved, but Microsoft states outright that the data shown is a sample of overall citation activity, not a full log. Intents and Topics give you the surrounding category and theme, which is closer to a real answer than the query list alone.

Why do Amazon and Walmart show up so often in Copilot answers?

Ahrefs’ analysis of more than 3 million US queries puts Amazon at 14.4% of all citations and Walmart at 10.0%, both with Domain Ratings in the 90s, and the study attributes it directly to Microsoft pushing shopping and checkout into Copilot chats. On broad commercial queries, that share is close to structural.

Get eligible everywhere, then get quoted

AEO Optimization fixes the crawl access and structure problems that keep you out of Copilot and every other answer engine, then restructures your pages around the intents Bing and Google both now report.

AEO Optimization, $445
95/100
AI Search ReadyOur own site is built to the exact standard we sell, so AI engines can find, read and recommend it. Verified by Faro, built by the same team.
See the proof →