Citation Share is the AI Performance metric Bing Webmaster Tools added in June 2026 that shows what percentage of all citations for a specific grounding query point to your site, turning a raw citation count into a share of visibility you can actually benchmark over time. It shipped alongside three companion features, Intents, Topics, and Compare, as an expansion of the AI Performance report Microsoft first launched in February 2026, and together they form the most detailed first-party AI citation dashboard any major search or AI provider has released to date.
Before this update, the AI Performance report told you that a page had been cited, and roughly how often, but not how that number compared to anything else. A count of forty citations for a query could mean you dominate every AI-generated answer on that topic, or that you are one voice among sixty competing sources. Citation Share closes that gap, and the three features released alongside it make the underlying data usable at scale instead of one grounding query at a time.
What shipped in the June 2026 update
Microsoft rolled out four expanded capabilities inside the existing AI Performance tab in Bing Webmaster Tools, live in preview worldwide from 16 June 2026: Intents, Topics, Citation Share, and Compare. None require a new integration or an extra verification step. If your site is already verified in Bing Webmaster Tools, all four appear automatically once your property has accumulated enough AI citation activity to populate them.
Intents
Intents classifies the grounding queries behind your AI citations into categories such as Informational, Commercial, Navigational, Learn and Solve, Research, Creation, and Local. Grounding queries are the underlying searches an AI system runs to retrieve sources before generating an answer, and a single user prompt can trigger several of them. Without Intents, you see a flat list of grounding queries with no way to tell whether your citations skew toward top-of-funnel informational traffic or bottom-of-funnel commercial research. With it, you can check whether your AI visibility actually matches the stage of the buyer journey you are trying to win.
Topics
Topics groups related grounding queries into broader thematic clusters, the same way a modern retrieval system organises information internally rather than treating every query as an isolated string. A page might be cited for a dozen differently worded grounding queries that all belong to one underlying topic; Topics rolls them up so you see concentration and gaps at the theme level instead of working through a long tail of near-duplicate phrasing one row at a time.
Citation Share
Citation Share is the headline addition. For a given grounding query, it reports the percentage of all citations that went to your domain rather than the raw count alone. If ten sources were cited across the answers Bing generated for a query and three of them were yours, Citation Share reports 30%. Microsoft has been explicit that it is an observational metric: it does not name competitor domains, does not represent click-through or traffic share, and does not assign a quality score to the content. What it gives you is a relative read on how much of the available citation space you are capturing for a topic you already care about.
Citation Share tells you your slice of the citation pie for a query, not who is eating the rest of it. Treat a low share as a prompt to test that query manually across ChatGPT, Perplexity, and Copilot to see which competitors are actually showing up, before assuming schema or content depth is the fix.
Compare
Compare lets you overlay two time periods on the same chart, so a spike or a decline in citation activity is visible against a defined baseline instead of inferred from memory. This is the feature that turns the other three from a snapshot into a measurement tool: publish a content change, wait two to four weeks for the retrieval layer to catch up, then use Compare to check whether Citation Share, citation count, or Topic-level coverage actually moved.
How to read Citation Share without over-interpreting it
The temptation with any new percentage-based metric is to treat it like a competitive scoreboard, chasing 100% the way you might chase a rank 1 organic position. Citation Share was not designed for that. A handful of caveats keep it useful instead of misleading:
- It is query-specific, not site-wide. A high Citation Share on one grounding query says nothing about your visibility on a related query with different phrasing; check the Topics rollup for the broader pattern instead.
- It excludes competitor identity by design. You will not see which domains make up the remaining share, only your own slice, so pair it with manual prompt testing in ChatGPT, Claude, Perplexity, and Copilot when you need to know who else is showing up.
- It is not a traffic proxy. A citation, unlike a click, does not guarantee a visit; a 40% Citation Share on a query that never sends a session is a visibility win with no commercial value yet, the same caution that applies to Google's AI Overviews impression data.
- It only covers Microsoft's surfaces. Citation Share reflects Copilot, Bing, and select partner AI experiences that draw on Bing's index, not ChatGPT, Claude, Gemini, or Perplexity, each of which needs its own tracking method.
A practical routine for using the new reports
The four features are most useful as a repeatable measurement loop rather than a dashboard you check once and forget:
- Record a baseline. Before making any content change, note your current Citation Share, citation count, and Topic distribution for the grounding queries that matter most to your business.
- Change one variable at a time. Rewrite a section, add FAQPage schema, or restructure a comparison table, but isolate the change so a later shift in Citation Share has a clear cause.
- Wait two to four weeks. Bing's retrieval and citation data needs time to reflect a content change; checking daily just adds noise to a metric that moves slowly by nature.
- Use Compare against your baseline period. Look for movement in Citation Share and citation count together, not either one in isolation, since a rising share of a shrinking query volume can still mean falling absolute visibility.
- Cross-reference Intents and Topics. If Citation Share climbs only on Navigational queries where you were already the obvious answer, the win is smaller than the same lift on Commercial or Research queries you are actively trying to earn.
Where this fits in a wider AEO measurement stack
Citation Share is the closest thing to a native competitive signal any AI provider has shipped, but it only covers one engine. Google's own AI performance reports, launched the same June 2026, go the other direction: they report AI Overviews and AI Mode impressions with no query-level detail and no percentage-based comparison at all. Reading the two side by side tells you more than either alone, since one exposes relative citation share with no query breakdown and the other exposes query-adjacent impressions with no share. Neither covers ChatGPT, Claude, or Perplexity, which is why manual AI citation gap analysis against named competitors, and GA4 referral tracking for the traffic these citations actually produce, remain the layers that fill in what neither native dashboard can see. If you have not built out the rest of the AEO measurement stack yet, Citation Share is a strong argument to prioritise it: a share number with nothing to compare it against is just a number.
If you already work on Copilot visibility, verifying your site in Bing Webmaster Tools is not optional groundwork, it is the only way to see any of this data. See how to optimise for Microsoft Copilot for the crawler access and schema fundamentals these reports assume are already in place.