Microsoft Adds AI Search Reporting to Bing. Is ChatGPT Next?

Are you struggling to measure how your brand performs in AI search? You’re not imagining the gap.

While marketers have been able to track organic rankings, paid impressions, and click-through rates for years, AI-generated answers have remained a black box — until now.

Microsoft has quietly dropped one of the most significant announcements in search marketing this year: AI performance reporting is now live inside Bing Webmaster Tools. It’s not Google. It’s not ChatGPT. But that’s almost beside the point.

For marketing leaders at businesses where decisions are made on data, this is the opening of a door that has been firmly shut. And the implications stretch far beyond Bing itself.

Executive Summary

• Microsoft has added AI performance tracking to Bing Webmaster Tools, covering Microsoft Copilot and Bing’s AI-generated summaries — its equivalent of Google’s AI Overviews.
• The dashboard includes total citations, average cited pages, grounding queries, visibility trends over time, and page-level citation activity.
• Grounding queries are a new metric showing the key phrases AI used when retrieving content — similar to keywords in Google Search Console, but native to AI search behaviour.
• This is the first inbound AI reporting of its kind: unlike existing tools that require you to pre-input queries, Bing’s dashboard shows what people are actually searching for in Copilot.
• The data is particularly valuable for B2B marketers, given the prevalence of Microsoft tools in corporate environments.
• Whether ChatGPT or Google will follow is uncertain — but the pressure is now on.
• If Bing traffic is thin, keyword research tools, AI search trackers, Reddit, sales calls, and customer conversations can help you build a proxy picture of AI search demand.

What Has Microsoft Actually Released?

The announcement centres on a new AI performance dashboard inside Bing Webmaster Tools. It’s not a complete overhaul — but what’s there is genuinely useful, and in some ways unprecedented.

Charlie Marchant, CEO of Exposure Ninja, describes it as “a pretty huge revelation.” The dashboard currently surfaces five key data points:

• Total citations — how often your content is referenced as a source in AI-generated answers
• Average cited pages — which pages are being pulled into AI responses most frequently
• Grounding queries — the key phrases AI used when retrieving content that was then cited in an answer
• Visibility trends over time — how your citation activity is changing across the site as a whole
• Page-level citation activity — which specific URLs are appearing inside AI-generated answers

The reporting covers both Microsoft Copilot and Bing’s AI-generated summaries — Bing’s version of Google’s AI Overviews. There’s also a reference to “partners,” though who exactly qualifies as a partner remains unclear. For now, don’t assume that includes ChatGPT, even given Microsoft’s existing relationship with OpenAI.

What the dashboard does not give you is full conversational detail. You won’t see the complete question a user typed. But the grounding queries offer something arguably more useful for strategy — and that’s worth understanding in depth.

Download Our 2025 State of AI Search Report

See how AI is reshaping search, which brands are already winning, and 2026 predictions for your AI Search strategy.

Download my free copy

What Are Grounding Queries — and Why Should You Care?

If “grounding queries” is new vocabulary for you, you’re in good company. The term needs a proper explanation before its significance lands.

When someone uses an AI chatbot — whether that’s Microsoft Copilot, ChatGPT, or Google’s AI Mode — they often type long, conversational questions. Recent data cited by Charlie puts the average AI prompt at around 60 words. That’s not a keyword; it’s a paragraph.

What happens next, in the background, is that the AI distils that prompt down into the core key phrase or topic it actually needs to search for. That distilled phrase — the one the AI used to go and retrieve content — is what Microsoft is calling a grounding query.

As Charlie explains: “To me this feels very similar to what we see in Search Console when we are looking at keywords and key phrases. I think they’ve used the phrase ‘grounding queries’ just to differentiate it — because some of these key phrases may be slightly different in length and are not going to match up exactly with data around Google search volumes.”

In practical terms: if someone asked Copilot, “I’m looking for a rugged commuter bike that can handle wet cobblestones in Edinburgh and fits in a car boot,” the AI might have retrieved your content by searching for something closer to “best commuter bikes.” That’s your grounding query. That’s what shows up in the dashboard.

For marketers, this is significant. It’s the first time any platform has reported on the actual queries triggering AI-cited content — without requiring you to guess what those queries might be in advance.

How Does This Differ From Existing AI Search Tracking Tools?

Tools like Peec AI, Scrunch AI, and Profound are genuinely excellent — and they’ve been the best available option for AI search visibility tracking up to this point. But they work in a fundamentally different direction to what Microsoft has now released.

With those platforms, you define the prompts and queries your business wants to track. You pre-programme the inputs, and the tool shows you visibility scores for those specific queries. You decide what matters; the tool measures it.

Bing’s AI performance dashboard works the other way around entirely. It shows you what people are actually searching for — unprompted, unfiltered. You’re not telling it what to look for. It’s telling you what’s already happening.

That’s a genuinely different kind of insight. As Charlie puts it: “A lot of people are like — Bing, that’s tiny. But what I think is interesting and exciting is the trend that this is showing for us.”

The real value isn’t just the Bing data itself. It’s having a dataset that reveals the shape of AI search demand in your category — data you can take beyond Bing and apply to your broader AI search strategy.

Why Is This Especially Valuable for B2B Marketers?

Bing’s share of overall search traffic is modest. In many industries, particularly for consumer-facing businesses in the UK, it barely registers. But B2B is a different story.

Microsoft’s suite — Outlook, Teams, Edge, Copilot — is deeply embedded in corporate environments. The people using Copilot are, disproportionately, professionals at work, on company devices, inside Microsoft ecosystems. That’s your procurement manager. Your IT director. Your CFO.

Charlie flags this directly: “Many businesses who sell to other businesses can find that Microsoft Suite is the type of suite they use — which means that if you particularly care about B2B search queries, this could be really insightful for you.”

For B2B marketing leaders running campaigns targeting senior decision-makers, the data from Bing Webmaster Tools has the potential to be genuinely representative of how your audience is using AI search — even if the raw volumes look underwhelming at first glance.

What Happens When AI Search Gets One Query and Turns It Into Many?

There’s an important technical concept sitting underneath all of this, and it’s worth understanding before you dive into the data: query fan-out.

When you submit a single question to an AI chatbot, that platform doesn’t just run one search. In the background, it fans the original query out into multiple sub-queries — sometimes 2 or 3, sometimes dozens, and in AI Mode, potentially hundreds — to build the most complete and contextual answer possible.

As Charlie explains: “What happens with our query is that even though we’ve put in just one keyword phrase — ‘best commuter bikes’ — what happens in the background is that AI can run 2, 3, 5, 10, hundreds of background queries. It takes that one query and fans it out into lots of other search queries it feels are relevant to give the context you’re looking for.”

So a single user prompt about commuter bikes might trigger background queries covering specific models, price ranges, tyre types, frame weights, and user reviews — none of which the user explicitly mentioned. The AI is trying to answer the deeper intent, not just the surface-level words.

This matters for your Bing data. The grounding queries you see in the dashboard may well represent some of those fanned-out sub-queries, not just the original prompt. Understanding that layered dynamic will help you interpret the data — and spot content opportunities you might otherwise overlook.

Will ChatGPT and Google Follow Microsoft’s Lead?

This is the question that every marketer wants answered. Bing is useful. Google Analytics with AI search data would be transformative. ChatGPT reporting with citation visibility would change the game entirely.

Charlie’s prediction is cautiously optimistic — with important caveats. On Google: “We would love Google to be doing some kind of similar reporting in Analytics.” On the pressure this creates: “This is going to put some pressure on OpenAI and Google to do something about AI analytics reporting.”

On ChatGPT specifically, though, she’s less bullish in the near term. OpenAI’s recent ad reporting — for the ads it’s been testing in the US — has been notably thin. Analytics depth doesn’t appear to be a priority for the platform right now.

The more likely near-term scenario is that this becomes a forcing function — a signal that sets expectations across the industry. Marketing teams will increasingly demand this data. The SEO community will be vocal about needing it. And over time, that pressure will build.

As Charlie puts it: “My hope is that the long game will be ChatGPT introducing something similar and Google introducing something similar — even if not exactly the same.”

What Should You Do If Your Bing Data Is Thin?

Not every business will find the Bing dashboard immediately populated with actionable data. If you’re in a sector with low Bing traffic, or you’re in a country where Bing usage is minimal, it might take time to build a meaningful dataset — or the volume might never reach the level of Google Search Console.

That’s fine. The answer isn’t to ignore it; it’s to use it alongside a broader intelligence-gathering approach. Here’s what that looks like in practice:

Set up Bing Webmaster Tools if you haven’t already. It takes around 48 hours to populate once set up, and if you already have Google Search Console configured, you can transfer the connection across directly. Even a small amount of data is better than none — and the setup cost is minimal.

Anchor your strategy in keyword research first. Tools like Semrush still provide the most reliable data on actual search volumes. Use that existing keyword intelligence to model what’s likely being searched in AI platforms. The overlap between traditional search intent and AI prompt intent is significant, particularly in competitive categories.

Use a dedicated AI search tracking tool. Peec AI offers a one-week free trial and is worth experimenting with. The key is going in with a clear list of prompts informed by your keyword research — you’ll get much more out of these platforms if you’ve done the groundwork first.

Mine forums and communities for real-world AI prompts. Dale Davies, Head of Marketing at Exposure Ninja, recommends a practical approach: “I’ll just open Quora and Reddit and see what kind of questions people are asking. Often a subreddit will have an FAQ area of questions that come up a lot — you can just compile those and add them into whichever tracking platform you use.”

Ask your customers what they searched for. Charlie highlights that this is particularly effective for regional and local businesses: “If you receive leads, ask them what they searched for. They’ll most likely remember something, particularly if it was terms like ‘best cybersecurity companies in London’ — they’re going to remember they roughly put that into ChatGPT.”

Analyse your sales call recordings. If you use a call recording and transcription platform such as Grain, you can create a folder of recent sales calls and use an AI tool to review them for language trends — specifically, how customers describe their problems and what they were searching for before they found you. This is a rich, underused source of AI search intelligence.

Review your live chat transcripts. The questions customers ask in live chat often mirror the questions they put into AI search. Reviewing those conversations systematically can surface the prompts your audience is using before they ever reach your website.

Next Steps

This week:

• Log into Bing Webmaster Tools and check whether it’s set up for your website. If not, connect it via your existing Google Search Console account — the process takes minutes.
• Locate the new AI performance dashboard and note what data is already populating. Even limited early data is worth reviewing.
• Flag this development to your wider marketing team. The grounding queries metric in particular deserves attention from anyone involved in content strategy or AI search optimisation.

Next 30 days:

• Cross-reference your grounding query data with your existing keyword research to identify any gaps or surprises — queries you’re appearing for in Copilot that you weren’t targeting, or notable absences.
• Set up or review your AI search tracking in a tool like Peec AI, using your keyword data as the foundation for your prompt list.
• Begin collecting AI search intelligence directly from customers — add a prompt to your sales process, review recent live chat transcripts, and identify your most active forums and communities.

Next 90 days:

• Build a consolidated view of your AI search visibility across Bing Webmaster Tools, your chosen tracking platform, and any qualitative data gathered from customers and communities.
• Use that picture to inform content strategy — identifying the question-based and intent-driven topics where your content can secure AI citations across Copilot, ChatGPT, Gemini, and Perplexity.
• Monitor whether Google or OpenAI announce any comparable reporting capability. If they do, the framework you’ve built will give you a significant head start.

In Conclusion

Microsoft’s AI performance dashboard in Bing Webmaster Tools is not a revolution in analytics. It’s a modest but meaningful first. And its significance lies less in the data it delivers today than in the direction it signals for tomorrow.

For the first time, marketers have a native reporting tool that shows them how AI search is actually interacting with their content — without having to pre-programme every query in advance. The grounding queries metric alone is worth paying attention to: it’s a glimpse into the AI-search behaviour of your audience, unmediated by your own assumptions.

The businesses that will be best positioned when Google and ChatGPT eventually follow suit are the ones building their AI search intelligence infrastructure now. That means setting up Bing Webmaster Tools today, combining it with keyword research and AI tracking tools, and gathering qualitative data from the customers who are already finding you through AI search.

The black box is starting to open. Get in before it’s fully ajar.

If you’d like an independent review of how your brand currently performs in AI search — and a clear roadmap for improving your visibility — request a free marketing review from Exposure Ninja.

Is your marketing underperforming?

Request a website and marketing review and our team will tell you how to improve your marketing.

Request a Free Review

*Disclosure: This page contains referral links. If you purchase through these links, we may receive a commission at no additional cost to you.