Bing Webmaster Tools Adds AI Performance Reporting: What ‘Citations’ Mean (and How to Use Them)

Bing Webmaster Tools Adds AI Performance Reporting: What ‘Citations’ Mean (and How to Use Them)

Microsoft has launched a new AI Performance report inside Bing Webmaster Tools. It’s positioned as an early step toward Generative Engine Optimization (GEO): instead of measuring only blue-link clicks, it helps publishers understand when their pages are cited as sources in AI-generated answers.

If you care about search traffic, this matters for a simple reason: AI answers are becoming a discovery surface. And on those surfaces, visibility often shows up as a citation rather than a click.

Primary sources:

• •Microsoft Bing Webmaster Blog — “Introducing AI Performance in Bing Webmaster Tools Public Preview”

https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

• •Search Engine Roundtable — “Bing Webmaster Tools Rolls Out AI Performance Report”

https://www.seroundtable.com/bing-webmaster-tools-ai-performance-report-40911.html

• •Microsoft Advertising Blog — “Optimizing Your Content for Inclusion in AI Search Answers”

https://about.ads.microsoft.com/en/blog/post/october-2025/optimizing-your-content-for-inclusion-in-ai-search-answers

TL;DR

• •Bing’s new AI Performance report shows how often your content is cited across Microsoft Copilot, AI-generated summaries in Bing, and “select partner integrations.” (Microsoft)
• •The core metric is Total Citations (how often your pages appear as sources in AI answers), plus Average Cited Pages, Grounding queries, and Page-level citation activity. (Microsoft)
• •This is not a replacement for classic search performance reporting. It does not tell you how much traffic you got from those AI answers, and Microsoft is explicit that citations don’t indicate ranking or placement. (Microsoft)
• •Practically: treat this as a signal for what AI systems are already using (and what they’re ignoring), then optimize the pages that should be cited.

What Bing’s AI Performance report measures

Microsoft frames AI Performance as an extension of Webmaster Tools’ existing role: crawl/index/technical health plus visibility reporting—now extended to AI-generated answers.

According to Microsoft, AI Performance reporting shows where and how publisher content is referenced as a source across AI experiences, including:

• •Microsoft Copilot
• •AI-generated summaries in Bing
• •Select partner integrations

Source: https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

The dashboard metrics (in plain English)

Microsoft lists five main components:

1) Total Citations

> The total number of citations displayed as sources in AI-generated answers during the selected time frame.

Translation: if Bing/Copilot is showing your URL as a source link inside a generative answer, it counts. Microsoft notes this is about frequency of citation, not placement.

2) Average Cited Pages

> Average number of unique pages from your site cited per day.

Translation: breadth of coverage—how many different URLs are being used, not just one winner page.

3) Grounding queries

> Key phrases the AI used when retrieving content that was referenced.

Translation: what the system “thinks” it’s searching for when it decides your page is a good reference.

4) Page-level citation activity

> Citation counts for specific URLs.

Translation: your internal leaderboard of “pages AI answers like to cite.”

5) Visibility trends over time

Translation: whether citations are increasing or declining for your site.

Source: https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

What it does NOT measure (and why you should care)

It’s tempting to interpret “citations” as the new “clicks.” Don’t.

Microsoft makes several important caveats:

• •Citations do not indicate ranking, authority, or role within any individual answer.
• •Metrics are aggregated across supported AI surfaces.
• •Grounding queries are a sample, and Microsoft says it will be refined as more data is processed.

Source: https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

Search Engine Roundtable also notes there’s no click data in this report.

Source: https://www.seroundtable.com/bing-webmaster-tools-ai-performance-report-40911.html

Why this limitation is still useful

Even without click data, citations can still be valuable because they answer questions SEOs have been guessing at:

• •Which of my pages are actually being used as “evidence” inside AI answers?
• •What topics does Bing/Copilot consider my site relevant for?
• •When citations drop, did it coincide with content changes, technical issues, or competitors improving coverage?

In other words: citations are a visibility signal—not a revenue metric.

How to use AI Performance reporting as a practical workflow

Here’s a conservative, low-risk way to turn the report into action.

Step 1: Identify “should be cited” pages that aren’t

Start with two lists:

• •Pages you expect to be cited (your best evergreen guides, definitions, comparisons, “how-to” pages)
• •Pages Bing says are being cited (page-level citation activity)

The gap is the opportunity.

Step 2: Make cited pages easier to “parse” and reuse

Microsoft’s guidance on AI inclusion repeatedly emphasizes structure and clarity—because AI answers often assemble content from pieces.

Practical edits that usually help:

• •Tighten the title + H1 to match intent
• •Add scannable H2/H3 sections that each answer one question
• •Include a short TL;DR that’s accurate and specific
• •Use bullets and tables for comparisons and steps
• •Add a small FAQ section where it genuinely helps

Source: https://about.ads.microsoft.com/en/blog/post/october-2025/optimizing-your-content-for-inclusion-in-ai-search-answers

Step 3: Support claims with sources (and keep pages updated)

Microsoft explicitly recommends supporting claims with evidence and keeping content fresh and accurate.

That aligns with how many AI answer systems behave: they prefer sources that are clear, current, and unambiguous.

Source: https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

Step 4: Use “grounding queries” to expand coverage

If grounding queries show you’re being retrieved for a topic cluster, consider adding:

• •a dedicated explainer page for the core concept
• •supporting sub-pages (examples, implementation, pitfalls, checklist)
• •internal links between them

This is standard SEO, but now you have a new lens: what the AI system is already using.

Step 5: Keep content fresh (and make changes discoverable)

Microsoft calls out IndexNow as a way to notify participating search engines when content is updated.

Even if you don’t implement IndexNow immediately, the underlying idea is straightforward:

• •update the pages that matter
• •ensure they’re crawlable
• •ensure the latest version gets discovered

Source: https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview

What this means for “GEO” without hype

The AI Performance report is not magic, but it’s a real step toward first-party measurement in the AI search era.

A grounded way to think about it:

• •SEO still matters (indexing, crawlability, site structure)
• •Content structure matters more because AI answers are assembled from fragments
• •Citations become a KPI for brand visibility and topical authority, even when traffic attribution is imperfect

If you already produce high-quality content, AI Performance can help you prioritize what to refresh, expand, or re-structure—based on what Bing’s AI systems actually cite.

Bottom line

Bing’s AI Performance reporting gives publishers a new type of visibility: how often they’re cited in AI answers across Copilot and Bing’s generative experiences.

Use it as a signal, not a scoreboard:

• •find the pages that earn citations
• •improve clarity and structure so those pages are easier to reuse accurately
• •build supporting content around the grounding queries that already show demand

It’s early, but it’s actionable.