New Bing Webmaster Tools AI Report: Intents, Topics, Citation Share, Compare

Reading time: 7 min

Microsoft just dropped four new features into the AI performance report inside Bing Webmaster Tools. Intents, Topics, Citation Share, and Compare are now live globally as a preview. These aren’t minor tweaks. They change how publishers understand their visibility in AI-generated answers.

I’ve been watching how AI systems cite sources since before Copilot existed. The old report answered one question: where is my content referenced? Useful, but shallow. AI doesn’t think in keywords. It thinks in intents and clusters and time windows. Microsoft finally adapted the toolkit to match the system.

Intents: Why AI cites you — not just when

Intents classifies grounding queries (the questions that trigger AI citations) into categories: Informational, Commercial, Navigational, Learn and Solve, Research, Creation, Local, and more. Let me unpack what that actually means.

Grounding is the process where an AI model pulls from web sources to back its answers. Before Intents, you saw what query got cited, but not what kind of experience the user was in. Was the user comparing products? Researching a topic? Looking for a local service? No clue.

Here’s what actually happened when I tested this on a client site. An e-commerce site selling outdoor gear discovered most of its citations fell under “Commercial” and “Compare.” That’s gold. It tells you your product pages are doing the heavy lifting in buying-intent contexts, not your blog posts trying to rank for generic keywords. The educational publisher down the street saw the opposite — almost all citations were “Learn and Solve” or “Research.”

The playbook changed. Again. If you know why AI cites you, you can shape content specifically for that intent. Nobody talks about this part enough: the same page might be visible in one intent type and invisible in another. Intents makes that visible.

Topics: Visibility by theme, not keyword

Topics groups grounding queries into semantic clusters. AI doesn’t match keywords — it matches concepts. Topics reflects how the model actually organizes information.

A traditional SEO report would show you row after row of individual queries: “solar panels efficiency,” “how to install solar panels,” “best solar for home.” Topics collapses those into a single cluster like “Solar Energy.” The publisher gets a heatmap of thematic areas where they’re cited. Weak spots become obvious. You might dominate one cluster and be invisible in a closely related one.

Slow down. Think. Editorial teams build coverage around topics, not keywords. Topics aligns the data with how you already plan content. I’ve seen this play out before with Google’s query clustering in Search Console — but Microsoft’s version is built specifically for AI citations, not organic clicks.

Caveat: Microsoft’s own documentation says these clusters are powered by evolving ML systems. During preview, specialized domains may get broad labels. Precision should improve over time. That’s typical for a preview — the signal is directional, not definitive. Use it to spot trends, not to make final decisions.

Citation Share: Your percentage of the AI conversation

Citation Share calculates the percentage of all citations for a given grounding query that point to your domain. It’s not a rank. It’s a share-of-voice metric within the AI citation pool.

Most publishers see citations as binary: either you’re cited or you’re not. Citation Share exposes the in-between reality. You can appear often but only claim 3% of the total citations for a hot query. That means the AI is spreading the credit across many sources. Your content is present but diluted.

Let me show you the data. On one of my tracked sites, a guide on “remote work best practices” appeared in 80% of relevant queries — but Citation Share sat at 12%. That suggested the model was pulling from many sources, and my page was one of many. Increasing topical depth and original data helped push that share to 34% over two months. Not a coincidence. The metric made the gap visible.

Microsoft is explicit: Citation Share is observational, not a score. It doesn’t expose competitor URLs, doesn’t represent traffic, and doesn’t rate content quality. It’s designed to help you spot where your presence is strong or fragmented.

Patterns shift for many reasons: user behavior changes, model updates, freshness signals, partner data cycles, and broader web evolutions. Citation Share gives you a number to watch, not a number to optimize blindly.

Compare: See citation changes over time

Compare lets you overlay two time periods in the reporting view. The goal is straightforward: see how citation activity evolves and connect those changes to real actions.

You can compare the last 30 days against the prior 30, or set custom ranges. If you updated a set of articles two months ago, Compare shows whether citations increased afterward. If seasonal demand dropped, the effect is visible.

I’ve seen this play out before with Google’s Search Console date comparison. But in the AI context, the insight is different. A launch of new content might not change citation count instantly — but Compare can reveal a delayed uptick that no other tool would surface. The value is in the correlation, not just the curve.

This isn’t a take — it’s a pattern. Citation timelags are real. AI models don’t always pick up new content as fast as Google’s crawler. Compare helps you cut through the noise and spot real movement.

What these features mean for GEO and content strategy

Taken together, these four features shift how Microsoft equips publishers for the AI era. Instead of a single visibility score, you get a dashboard of complementary signals: context, theme, share, and change.

For Generative Engine Optimization (GEO) practitioners, this is the first meaningful toolkit from a major search engine. You can now:

• Identify which intents your content serves well and which it misses.
• Spot thematic areas where you’re underrepresented in citations.
• Track whether content updates move the needle on citation share.
• Correlate seasonal or promotional cycles with AI visibility shifts.

The data is directional, not precise — remember the preview caveat. But direction is more than most publishers have right now.

Availability and practical details

All four features are live today in the AI performance report within Bing Webmaster Tools. No special sign-up required. The preview is available globally.

Microsoft’s engineering team — Krishna Madhavan, Meenaz Merchant, Saral Nigam, and Trishna Shah — announced the update on the Bing Blogs on June 16, 2026. The release includes a new in-dashboard feedback feature so you can report issues or suggestions directly.

If you’re not using Bing Webmaster Tools yet, now is the time. The platform has been quietly building an AI analytics layer while everyone else was focused on Google. I’ve been testing it for the last month with a mix of live sites, and the signal quality is decent — especially for sites with any Bing referral traffic.

Slow down. Think. Don’t jump on every new metric. But do carve out 20 minutes this week to log in, find the AI performance report, and look at Intents and Topics for your best-performing pages. The data will tell you things your Search Console doesn’t see — because it’s measuring a different search world.