Bing Webmaster Tools have made AI visibility measurable since February 2026 with a dedicated AI Performance report. As of June 3, 2026, Google is following with its own dedicated reports.
- Bing shows how often and where your content is cited in AI answers from Microsoft Copilot and Bing - with Citations, Grounding Queries and page-level citation activity.
- June 2026 update: Bing expanded the report with four analysis dimensions - Intents, Topics, Citation Share and Compare (preview, rolled out globally on June 16, 2026).
- Google’s new GenAI performance reports (June 3, 2026) show impressions only, according to Google - on data depth, Bing remains ahead.
- The AI Performance report is the first real step toward Generative Engine Optimization (GEO) with reliable first-party data - set it up now.
Let’s be honest: do you actually know if your content is being cited by AI systems? When users ask Microsoft Copilot or Bing a question, the AI increasingly generates direct answers - synthesized from various web sources. Your content might be referenced as a source. Or it might not. Until now, this was a complete black box.
On February 10, 2026, Microsoft broke that box open. With the new AI Performance Dashboard in Bing Webmaster Tools, you can see for the first time how often your content is cited in AI-generated answers, which specific pages are referenced, and which search queries trigger those citations.
The good news: you don’t have to wait for Google. This data is available now - in Public Preview - providing a brand-new perspective on your AI visibility. And yes, I activated the dashboard for seo-kreativ.de immediately - with surprising results.
The numbers speak for themselves: 829 citations in 30 days, with an average of 6 pages per day referenced as sources in AI answers. This is for a niche SEO blog with German content that was barely six months old at the time of this analysis. Particularly interesting: the top grounding queries were “tl dr meaning” (107 citations) and “tl;dr” (67 citations) - a blog post I wrote primarily as a glossary explanation with an SEO twist, not as content specifically optimized for AI. Third place was “google core update december 2025” with 20 citations. This shows that AI pulls content it deems thematically authoritative - not necessarily content engineered for citations.
What is the AI Performance Report?
The AI Performance report is a new, standalone section within Bing Webmaster Tools. It doesn’t track clicks or rankings in the traditional sense; instead, it tracks something fundamentally different: if and how often your content is used as a source in AI-generated responses.
The data is aggregated across several Microsoft AI interfaces, including AI summaries in Bing Search, Microsoft Copilot, and selected partner integrations.
Microsoft itself describes the launch as an “early step toward Generative Engine Optimization (GEO) tooling” - the first official step in making GEO measurable. This is the point: we aren’t talking about a nice-to-have extra. We are talking about the first reliable data source for a discipline that previously relied entirely on guesswork.
Why it matters
Imagine you run a guide-style blog. Your content ranks well in traditional search. But what happens when a user asks Copilot that same question instead of typing it into a search bar? Will your content be cited? Or will the traffic go to another source?
The AI Performance report answers exactly that. And the answer is sometimes surprising: for seo-kreativ.de, it wasn’t the complex deep-dive analyses cited most often, but a simple post about the meaning of “TL;DR” - cited 107 times in 30 days. AI follows its own logic. And if you thought Bing was irrelevant - remember: Microsoft Copilot is already integrated into Windows, Office, and Edge. Its user base is growing quietly but steadily - completely independent of classic Bing search market share.
The New Metrics Explained
The dashboard introduces four core metrics that differ fundamentally from classic SEO KPIs. This isn’t about rankings or CTR - it’s about citation frequency and thematic relevance.
| Metric | What it shows | What to watch for |
|---|---|---|
| Total Citations | Total number of times your site was cited as a source in AI answers during the selected period | Watch the overall trend - is your AI visibility rising or falling? |
| Average Cited Pages | The average number of unique URLs per day referenced in AI answers | The breadth of your AI presence - how many of your pages are used at all? |
| Grounding Queries | The search phrases the AI used when retrieving your content | Shows which topics the AI considers your content relevant for |
| Page-level Citation Activity | Citation counts per URL - which pages are referenced most often | Identify your AI power pages and find pages with optimization potential |
Grounding Queries: perhaps the most exciting metric
Grounding Queries deserve special attention. They don’t show user search queries; they show the phrases the AI system used internally to retrieve your content. This is a subtle but vital distinction.
In my assessment, Grounding Queries are the most exciting data point in the entire dashboard. They give you insight into the “mindset” of the AI system - showing the thematic connections the AI makes between user questions and your content.
What my own Grounding Queries reveal
A look at my own data makes this tangible (as of June 2026): my top grounding query is now “Google AI Mode vs AI Overviews” with 109 citations - a head-to-head comparison piece the AI apparently treats as one of the clearest German-language sources for that distinction. Close behind are “google indexierung” with 83 and my TL;DR glossary post, which across several query variants (“tl dr meaning”, “tl;dr”, “tl dr bedeutung”) adds up to well over a hundred citations combined. These figures are only a sample and shift constantly - which is exactly why a regular look at the dashboard pays off.
This changes how we think about keyword research. It’s no longer just about search volume and ranking positions, but about the question: for which concepts does the AI consider your content citation-worthy? Grounding Queries are the AI equivalent of search queries in Search Console - but they show you something fundamentally different: not what people are searching for, but what the machine reaches for when it needs to build an answer.
June 2026 update: Intents, Topics, Citation Share and Compare
On June 16, 2026, Microsoft sharpened the report and rolled out four new analysis dimensions as a global preview. They all hang off the grounding queries and answer the question the original dashboard left open: not just what you get cited for, but in which context - and how big your share of it is.
| New dimension | What it shows | What you use it for |
|---|---|---|
| Intents | Classifies each grounding query by intent - e.g. Informational, Comparison, Commercial, Research or Local | See whether the AI pulls you in for knowledge questions or for purchase-oriented comparisons |
| Topics | Groups related grounding queries into thematic clusters instead of single keywords | See which broader topic areas your authority carries - not just individual phrases |
| Citation Share | Your percentage share of the total citation space for a query - relative rather than absolute visibility | Gauge how dominant you are for a query compared to all other cited sources |
| Compare | Direct comparison of two time periods inside the report view | Connect changes to your own content updates or seasonality |
For me, Citation Share is the most important of the four. Total citations tell you how often you get cited - but not whether you dominate the answer or are just one footnote among ten sources. That is exactly the gap this column closes.
My own data makes the value obvious right away: my top query “Google AI Mode vs AI Overviews” is classified as intent Comparison, sits under the topic “AI Tools & Platforms” and reaches a Citation Share of around 38 percent - so for that question roughly every third AI citation comes from my site. For “E-E-A-T Bedeutung AEO” it is a good 18 percent, for “tl dr bedeutung” just under 13 percent. That is a different quality of information than a bare citation count: I do not just see that I appear, but how heavily my voice is weighted in the AI answer space.
One caveat remains even after the update: the new dimensions run as a preview, and click data is still missing. According to Microsoft the data is not yet available via the API - that is still on the roadmap.
Google vs. Bing: Who Provides Better AI Data?
This is where it gets really interesting - and frankly, a bit frustrating if you primarily optimize for Google. The comparison clearly favors Bing.
| Feature | Bing Webmaster Tools | Google Search Console |
|---|---|---|
| Dedicated AI report | Yes - standalone AI Performance Dashboard | Yes - dedicated GenAI report, worldwide since 31 August 2026 |
| Citation data | Yes - Total Citations, page-level activity | No - no separate citation metrics |
| Grounding Queries | Yes - AI-internal search phrases visible | No |
| Separate AI traffic | Yes - own section | Yes - dedicated view, impressions only |
| AI click data | Not yet (announced) | Not in the GenAI report, only aggregated in the web report |
Google officially documents how AI Overviews and AI Mode are counted in Search Console. A dedicated view for it has existed since 3 June 2026, the GenAI performance report. It reports impressions, pages, countries, devices and the time series. Click data is not in it so far, that stays aggregated in the standard web report. An equivalent to Bing’s Citations or Grounding Queries is missing there too.
SEO expert Aleyda Solis aptly stated on X that this was the first official AI search visibility tool ever. Koray Tuğberk Gübür went even further, noting that Bing Webmaster Tools was once again superior to Google Search Console in terms of transparency.
In my assessment Google had to follow suit. It did. The SEO community had been asking for separate AI data in Search Console for months. John Mueller had still exposed a supposed screenshot of an AI Overviews filter in GSC as a fake in September 2025. Nine months later the GenAI performance report is here. The gap remains all the same: Microsoft provides Grounding Queries, a metric for which Google still has no equivalent.
How to Optimize Your Content for AI Citations
The data from the AI Performance dashboard is just the start. The real work begins afterward: preparing your content so AI systems prefer to use it as a source.
In the official announcement blog post, Microsoft provided specific tips that align closely with what we know from E-E-A-T optimization.
Create structure and clarity
AI systems prefer to extract information from clearly structured content. Clear heading hierarchies (H2/H3), tables, and FAQ sections make it easy for AI to identify and correctly reference relevant passages.
If you regularly implement structured data, you already have an advantage. The same logic applies to AI citations: the clearer the structure, the more likely the reference.
Thematic depth over keyword stuffing
Pages cited for specific Grounding Queries usually demonstrate high thematic authority. This is no accident. AI systems prefer content that covers a topic comprehensively and with genuine expertise - not pages that superficially check off keywords.
Entity consistency: speak the AI’s language
A point often missed in AI optimization discussions: AI systems don’t work with keywords, they work with entities - clearly identifiable concepts, people, products, or topics. If you switch between “Google Core Update”, “Google algorithm update”, and “the latest big Google update” on one page, you create unnecessary ambiguity for the AI. It first has to decode whether you mean the same concept or different ones.
My own Grounding Queries illustrate this: the AI retrieves my content via both “tl dr meaning” (107 citations) and “tl;dr” (67 citations) - two spellings of the same entity. My post covers both variants consistently, which is why it’s cited for both. The rule of thumb: use a clear primary name per concept and explicitly introduce common variants. Link related entities so the AI recognizes the thematic connection. This isn’t a new SEO principle - but in the context of AI retrieval it becomes a deciding factor, because the machine has no visual cues or layout context. It only understands what is linguistically unambiguous.
Ensure freshness
Microsoft explicitly recommends using IndexNow to immediately notify search engines of updates. AI systems favor current information. If your content is outdated, a more current source will be cited - it’s that simple.
Back up claims with evidence
Data, examples, and source citations increase trust in your content. When an AI has to choose between two similar sources, the one with better evidence wins. This aligns with everything we know about quality signals and trust.
Don’t forget technical performance
What is often overlooked: AI systems must generate responses in milliseconds. They cannot wait for slow servers. Bing emphasizes that page speed - especially a low Largest Contentful Paint (LCP) - is a factor in AI retrieval probability. If your page loads quickly, is cached, and has minimal redirects, your chance of being cited increases. So don’t just think about content, but also about the technical foundation behind it.
Local businesses: don’t forget Bing Places
For local service providers there is an additional lever: Bing Places for Business. Microsoft explicitly recommends keeping business info like address, hours, and contact data up to date there. When AI systems generate answers for location-based queries, they pull from this data. If you aren’t listed or maintained there, you will simply be ignored in local AI answers.
Checklist: optimizing AI visibility
| Step | Action |
|---|---|
| 1 | Verify your site in Bing Webmaster Tools (import from Google Search Console is available) |
| 2 | Open the AI Performance dashboard and check current citation data |
| 3 | Analyze Grounding Queries - for which topics are you used as a source? |
| 4 | Identify your most-cited pages and further enhance their content quality |
| 5 | Revise rarely cited but indexed pages: structure, depth, freshness |
| 6 | Check technical performance: optimize load times, keep LCP low, minimize redirects |
| 7 | For local businesses: set up Bing Places for Business and keep data current |
| 8 | Implement IndexNow to communicate content updates to Bing instantly |
| 9 | Review and document visibility trends in the dashboard monthly |
Setting Up AI Performance: Step by Step
If you aren’t using Bing Webmaster Tools yet, now is the perfect time. Setup is straightforward and takes less than five minutes.
Open the Bing Webmaster Tools and sign in with a Microsoft, Google, or Facebook account. Next, add your website. The fastest way is to import your property directly from Google Search Console. This carries over sitemaps and basic settings.
Once verified, you’ll find the AI Performance report as its own menu item in the left-hand navigation. It immediately shows the available data - provided your pages have already been cited in AI answers.
In addition, Bing has given Webmaster Tools a complete redesign. The new interface is much cleaner and more modern - and genuinely one of the most user-friendly webmaster tool interfaces I’ve seen in years. Anyone who knew the old Bing interface will notice the difference instantly.
Frequently Asked Questions (FAQ)
Does the AI Performance report show click data?
No, not currently. The report shows citation frequency, but no click-through data from AI answers yet. However, Microsoft has hinted that more metrics are coming. Fabrice Canel, Principal Product Manager at Microsoft Bing, stated on X: “It’s just a preview, you will get more in 2026.”
Is this worth it for markets where Bing has low market share?
Yes, for two reasons. First, Copilot usage via Windows and Office is growing steadily, regardless of Bing’s search market share. According to Microsoft Clarity data, AI-based referrals have increased massively in recent months and sometimes convert significantly better than classic organic traffic. Second, the dashboard provides general insights into your AI readiness that can be applied to Google AI Overviews and other systems.
What exactly are “Grounding Queries”?
Grounding Queries are the internal search phrases Bing’s AI system uses to retrieve relevant web content for an answer. They don’t necessarily match the user’s query but show how the AI interprets your topic area. Think of it as the AI equivalent of search queries in Search Console.
Can I use Bing Webmaster Tools without specifically optimizing for Bing?
Absolutely. You don’t need to change your SEO strategy. Simply verify your site and use the AI data as additional insights. Information about your AI visibility is valuable regardless of whether you actively target Bing.
How many citations are considered “good”?
This depends heavily on your niche and content volume. For reference: seo-kreativ.de, with around 200 indexed pages and a domain age of just over half a year, reaches 829 citations in 30 days with an average of 6 cited pages per day. The trend - is visibility rising or falling? - matters more than the absolute number. And which pages are cited? Often they aren’t the ones you expect.
How does the AI Performance report differ from the normal search reports?
The classic Search Performance report shows clicks, impressions, CTR, and position in the regular Bing search results. The AI Performance report, by contrast, shows only how often and which of your pages are cited as a source in AI-generated answers - a completely different data basis.
What do the new Intent, Topic and Citation Share columns show?
Since the June 16, 2026 update (preview), Bing enriches each grounding query with three data points: Intent classifies the purpose (e.g. Informational or Comparison), Topic groups related queries into thematic clusters, and Citation Share shows your percentage share of the citation space for a query. In addition, the Compare function lets you compare two time periods directly. So you no longer just see how often you are cited, but in what context and how dominantly.
Conclusion: The First Real GEO Data Point is Here
With the AI Performance Dashboard, Microsoft has delivered what the SEO industry has been demanding since the rise of AI search: reliable first-party data on AI visibility.
And the data delivers. 829 citations in 30 days for seo-kreativ.de show that even a German-language niche blog is regularly used as a source in AI answers. The Grounding Queries reveal surprises - the AI doesn’t necessarily cite the content you think is important, but what it deems thematically authoritative.
Of course, the report isn’t perfect yet. Click data is missing, the Grounding Queries are currently just a sample, and Bing naturally has less relevance in Germany than Google. But that doesn’t diminish the importance of this step. For the first time, you can measure whether your content appears in AI answers - and derive concrete optimization measures from it.
But be honest with yourself too: the report shows citations, not impact. As long as click data is missing, it remains open whether AI visibility actually delivers traffic and conversions - or whether the AI uses your content without ever sending a user to your site. That’s the uncomfortable question Microsoft hasn’t answered (yet). Still: the alternative is flying blind. And flying blind was never a good strategy.
Google had to follow suit - and as of June 3, 2026 it has, with its own GenAI performance reports. Until they match Bing’s data depth, you’re a step ahead with the Bing Webmaster Tools.
In twelve months, you’ll be glad you started today: AI search will keep gaining importance, Google will keep expanding its AI reports, and those collecting data now will have a real head start - with trends, benchmarks, and optimization experience that no competitor can easily replicate.
