Bing Webmaster Tools previews Citation Share for AI queries

What happened

Microsoft previewed four new AI reporting features for Bing Webmaster Tools at SEO Week in New York City. Krishna Madhavan, Principal Product Manager at Microsoft AI and Bing, showed the additions during a presentation, according to Search Engine Journal’s coverage.

The four features target the existing AI Performance dashboard:

• Citation Share would show the percentage of citations a site captures within a specific grounding query, sitting alongside the raw citation counts already in the dashboard.
• Grounding Query Intent would classify queries into 15 predefined intent labels. Screenshots shared by attendees on X show labels including Learning, Informational Search, Navigational, Research, Comparison, Planning, Conversational, and Content Filtered.
• Grounding Query Topic would group queries under topic labels, adding a second classification layer alongside intent.
• GEO-focused recommendations would surface guidance tied to AI visibility. The slide showed recommendation areas covering content structure, crawlability, indexing and canonicalization signals, structured data adoption, and structured data quality.

Microsoft has not published an official blog post about these features. The details come from attendee screenshots of the presentation.

Why it matters

The AI Performance dashboard launched in public preview in February 2026 and gave sites their first look at how often Copilot and Bing AI summaries cite their content. Microsoft expanded it in March with a feature mapping grounding queries to specific cited pages. Citation Share would add competitive context to those raw counts.

Knowing you received 12 citations for a query is useful. Knowing those 12 citations represent 80% of all citations for that query is more useful. The share metric tells you whether you dominate a grounding query or split visibility with competitors.

The intent and topic classifications address a real data problem. Grounding queries vary widely in phrasing, making trend analysis difficult. Grouping by intent and topic would let sites gauge visibility across categories rather than chasing individual query strings.

The GEO recommendations are the least defined of the four. The visible labels suggest the focus areas are familiar SEO fundamentals: crawlability, indexing, canonicalization, and structured data. Microsoft hasn’t explained how recommendations would be generated or triggered.

The timing is notable. Google has also started surfacing AI Mode traffic data in Search Console, though Google’s approach uses traditional impression and click metrics rather than citation-specific reporting. Bing’s Citation Share concept has no direct equivalent in Google’s tooling yet.

What to do

No action is needed right now. These are previews, not shipped features. Microsoft has not announced release dates for any of the four additions.

If you haven’t already, set up and verify your site in Bing Webmaster Tools. The AI Performance dashboard is already live and shows grounding query data. Familiarize yourself with the existing reports so you have baseline data when the new features roll out.

Watch for official announcements on the Bing Webmaster blog or Microsoft Advertising blog confirming scope and timing. Until then, treat the screenshots as directional, not final.

For sites already tracking AI citation performance, note the intent taxonomy. The 15 predefined labels suggest Microsoft is building a standardized classification system. Understanding which intent categories your content serves in AI results could shape content strategy once the labels ship.

Watch out for

Preview vs. reality. The features were shown in a conference presentation, not an official product announcement. Feature scope, naming, and availability could change before release. Do not build reporting workflows around details that may shift.

GEO recommendations may be generic. The visible recommendation areas (crawlability, structured data, canonicalization) overlap heavily with existing SEO best practices. Wait to see whether the actual recommendations are site-specific or boilerplate before adjusting priorities based on them.