Bing’s New AI Performance Report Just Changed How We Should Measure AI Visibility

For years, the way we measured success in search was simple and unchanging. You looked at rankings. You looked at impressions. You looked at clicks. You looked at click-through rate. Every dashboard, every report, every conversation about whether a piece of content was “working” came back to those four numbers.

Then AI search happened, and that whole measurement model started to crack. AI engines do not just rank your page and hand a user a link. They read your content, synthesize it with other sources, and hand the user an answer. The user might never click anything. They might never even know your name. But your content shaped what they were told – and the old four metrics have no way of capturing that.

Microsoft just gave the industry a real answer to this problem, and it is worth paying close attention to. Bing Webmaster Tools has expanded its AI Performance Report with four new capabilities: Intents, Topics, Citation Share, and Compare. On the surface this looks like a routine product update. Underneath it, it tells you exactly where AI search measurement is heading – and the detail that matters most is not any single feature. It is what Microsoft chose to build the entire report around.

They are not measuring rankings. They are measuring citations. That single decision says more about the future of search than almost anything else happening in this space right now.

What Bing Actually Launched

Let’s start with some quick context. Bing introduced the original AI Performance Report back in February 2026 – the first time any major search engine gave publishers an official window into how their content was being used inside AI-generated answers across Microsoft Copilot and Bing’s AI summaries. It was a genuinely significant first step, but a fairly blunt one. It told you how many times you had been cited and which pages were doing the citing. Useful, but limited – like a phone bill that only shows total minutes with no breakdown of who you actually called.

The update that rolled out on June 16, 2026 changes that completely. As Microsoft’s Krishna Madhavan explained when announcing it, these new capabilities are designed to help publishers understand why their content is being surfaced, which broader subject areas they are gaining visibility in, how their presence evolves relative to other cited sources, and how citation patterns change over time. That is a much more sophisticated set of questions than the original report was built to answer.

Breaking Down Each of the Four New Features

Intents – the why behind the citation

The Intents feature sorts the queries behind your citations into categories: Informational, Commercial, Navigational, Learn and Solve, Research, Local, and a few others. Instead of looking at citations one at a time and trying to guess at a pattern, you can now see the shape of your AI visibility in one view.

This matters more than it sounds. An e-commerce brand might discover that almost all of its citations come from commercial or comparison-style queries – a clear signal to lean further into that kind of content. An educational publisher might find the opposite, with most citations tied to research and learning intent, pointing toward a very different content strategy. Same dashboard, completely different story, depending on what your business actually does.

Topics – AI organizes by theme, not by keyword

The Topics feature groups related queries into broader thematic clusters. Microsoft’s own description of this is worth repeating directly, because it is unusually candid: it is designed to help publishers understand visibility in the same thematic structure that modern AI systems use to organize information.

That sentence is Microsoft confirming something out loud that a lot of the industry has only been speculating about. Traditional search engines organized the world by keyword. AI systems organize the world by topic and meaning. Bing building an entire reporting feature around topic clusters rather than individual search strings is the company publicly acknowledging that the old keyword-based model of measurement no longer reflects how its own AI systems actually think.

Citation Share – probably the most important number in this whole update

This is the feature worth sitting with the longest, because it might end up being the most consequential. Here is the plain version of how it works, straight from Microsoft’s own description: Citation Share shows what percentage of the total citations for a specific query your site captured. If a query generated ten total citations across every source involved, and three pointed to you, your Citation Share for that query is 30 percent.

The shift here is subtle but important. A raw citation count tells you that you got mentioned. Citation Share tells you how much of the available opportunity you actually captured relative to everyone else competing for the same query. It is the difference between knowing you scored points and knowing what percentage of the entire game you controlled.

One important clarification, since this detail is easy to get wrong: Microsoft has been explicit that Citation Share is an observational metric. It does not show named competitor domains, and it does not expose traffic data. You can see your own share of the citation space for a query, but you do not get a leaderboard naming exactly who else is in it. Think of it less as competitive intelligence and more as a self-assessment tool, at least in its current preview form.

Compare – finally, a way to know if your work is paying off

Compare lets you overlay one time period against another – the current 30 days against the prior 30, or any custom range you choose – so you can see whether your citation activity is actually trending up, down, or staying flat. It sounds almost too simple to be worth mentioning, but it solves a real, persistent problem. Without it, every number you look at is just a snapshot with no way to know if it represents progress or noise.

If you publish a major update, refresh a key page, or build out a new content cluster, Compare is what finally lets you close the loop and see whether that effort actually moved your citation numbers. That feedback loop has been one of the biggest gaps in GEO so far, and this is a real step toward closing it.

The Feature That Did Not Make the Cut

Here is a detail most of the immediate coverage skipped past, and it is worth pausing on. At SEO Week in late April, Microsoft had also previewed a fifth capability: GEO-focused recommendations, expected to surface specific guidance on crawlability, structured data, and indexing issues.

That feature did not ship in the June rollout, and there is no stated timeline for when, or if, it will arrive. My honest read is that this reflects something genuinely difficult about where this technology currently stands. Measuring and reporting citation activity is one challenge. Reliably translating that activity into specific, trustworthy, actionable recommendations is a much harder one, and Microsoft was evidently not ready to ship that part yet.

This is worth remembering anytime a platform or tool promises to tell you exactly what to fix. The current state of the art is genuinely good at showing you what is happening. It is far less mature at confidently telling you why, or precisely what to do about it. Treat bold claims to the contrary with a healthy amount of skepticism.

How This Compares to What Google Is Doing

Context matters here, and the comparison is genuinely informative. Google has been testing its own AI visibility reporting inside Search Console, but industry coverage at the time of this Bing update described Google’s version as arriving later and feeling somewhat rushed by comparison. Bing’s AI Performance Report first launched back in February, months ahead of anything comparable from Google.

This is a notable moment in the broader competitive landscape. For most of the last two decades, Google was the unquestioned center of gravity for search measurement. Bing moving first, and more comprehensively, on AI citation reporting is a real signal that AI search measurement is not simply going to be Google with an AI layer bolted on top. There is a genuine race underway, and right now Microsoft has a credible claim to leading it on this specific front.

That said, scope matters. These Bing features only cover Microsoft Copilot and Bing’s own AI experiences. They tell you nothing about your performance inside ChatGPT, Google’s AI Overviews, Perplexity, or any other AI surface. For most businesses, audiences are spread across several of these tools, which means Bing’s dashboard is one valuable data source among several you will need – not a complete picture by itself.

What to Keep in Mind Before You Get Too Excited

A few honest limitations are worth knowing before treating this data as gospel. Microsoft’s own documentation is fairly direct that the grounding queries shown in the report are sampled, not exhaustive, and are grouped into generalized phrases rather than exact individual prompts. The data also aggregates across all of Bing’s supported AI surfaces, so you cannot currently isolate Copilot activity specifically from Bing’s AI summaries or partner integrations.

And critically, a citation is not the same thing as a click, a ranking, or even prominence within an answer. Being cited tells you your content was used as a source. It does not tell you whether you were the first source mentioned, the most prominent one, or whether the person who saw that answer ever did anything as a result.

None of this makes the tool less valuable. It just means the right posture is the one any good analyst takes with early-stage data: genuinely useful for spotting trends over time, but not something to over-index on for granular, page-by-page decisions just yet. Microsoft itself has acknowledged that the quality and precision of these features will keep improving as more data flows through the system, which is a fair expectation for something this new.

What This Actually Means If You Run a Website or Manage Content

If you have not already verified your site in Bing Webmaster Tools, that is the first and easiest step, and it costs nothing. Even if most of your audience lives elsewhere, having this baseline data quietly accumulating in the background means you are not starting from zero whenever you do decide to dig in seriously.

Once you have access, the most useful exercise is not staring at your total citation count. It is studying the pages that are already earning citations and asking what they have in common. Are they your glossary or definition pages? Comparison content? Original research? How-to guides? That pattern tells you what Bing’s AI systems currently consider strong source material on your site – and gives you a strong hint about what kind of content elsewhere is likely to perform well too.

From there, export your grounding queries periodically and cluster them yourself by topic and intent – the same categories the new Intents and Topics features are now organizing automatically. Look specifically for gaps: topics where you have some citation activity, but clearly less than the depth of your content on that subject would suggest. That gap is usually a sign the content exists but is not structured clearly enough to be extracted easily by AI systems.

And once Citation Share and Compare are available on your account, use them the way they are genuinely meant to be used: as a feedback loop, not just a scoreboard. Make a substantial update to an important page, mark the date, and check back in thirty to sixty days to see what actually moved. This is one of the first times anyone outside Microsoft has had an official, first-party way to test that loop directly instead of guessing at it.

The Bigger Shift Hiding Inside a Product Update

Step back from the individual feature names for a moment, and the real story here is simple. For most of the past two decades, the entire measurement architecture of search marketing was built around one core idea: where do you rank, and how many people click. Every report, every dashboard, every KPI that mattered in SEO traced back to that idea in some form.

Microsoft just built an entire reporting suite that does not mention rankings once. It is built from the ground up around a different question entirely: when an AI system answers a question, how much of that answer space do you actually own. That is not a minor tweak to existing measurement. It is a fundamentally different way of thinking about what success even means in search.

Will Citation Share become the defining KPI for AI search visibility in the years ahead? Based on everything happening across this space right now – the growing emphasis on topical authority, on structured and extractable content, on earned credibility, on technical accessibility – it certainly looks like the industry is heading in exactly that direction, even if the specific tool or metric name continues to evolve as the space matures.

The people and brands who start getting comfortable with this way of thinking now – measuring presence and share of the citation space rather than just position and clicks – are going to have a real head start over everyone still waiting for the old playbook to come back.

Contributed by GuestPosts.biz