How to Track AI Visibility with Bing’s AI Performance Report

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.

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 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.

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 the seo-kreativ.de data makes this tangible: my top grounding query is “tl dr meaning” with 107 citations – a post I wrote as a glossary explanation with an SEO angle. The AI clearly views it as one of the most authoritative German sources for this term. At number 3 is “google core update december 2025” with 20 citations – here, the AI specifically pulls from my update analysis.

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.

This is where it gets really interesting – and frankly, a bit frustrating if you primarily optimize for Google. The comparison clearly favors Bing.

Google officially documents how AI Overviews and AI Mode are counted in Search Console. With the new GenAI performance reports there is now a dedicated impressions view as well – but no equivalent to Bing’s Citations or Grounding Queries.

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 – and it did, with its own GenAI performance reports as of June 3, 2026. Still, the real difference remains the data depth: Microsoft provides Grounding Queries, a metric for which Google has no equivalent.

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

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.

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.

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.