Bing’s AI Performance report: the free AI-citation data most B2B teams haven’t opened

Most B2B teams can tell you where they rank on Google. Very few can tell you how often an AI assistant cites their pages as a source. Bing now gives you that number for free, and the report has been sitting in Bing Webmaster Tools since February.

It matters because AI answers are where a growing share of buyer research starts, and until recently the only way to see your citations was to pay a third-party tool that samples AI answers from the outside. Bing’s AI Performance report is first-party data. It only covers Microsoft’s ecosystem, but it is the first time a major search engine has shown publishers how often AI answers cite them.

This article explains what the report shows, what it leaves out, and a 30-minute routine for turning it into content decisions. It’s written for CMOs, marketing leaders and founders who want AI visibility numbers they can defend.

What the report actually shows

The report lives in Bing Webmaster Tools and needs only a verified site. If you already use Google Search Console, you can import your sites instead of verifying them again. Bing defines the data as follows.

According to Practical Ecommerce’s coverage of the June update, each grounding query now also shows its intent (navigational, informational, commercial or comparison), a topic label, average unique pages cited per day and your citation share. You can sort on those fields, drill from a query to the pages cited, and export the full report to a spreadsheet.

What it leaves out

The gaps are as useful to know as the data, because they decide how you can report on it.

• No clicks. The report shows visibility only, so you cannot tell whether a citation led to a visit. Per Momentic, this is the main limit.
• Bing and Copilot only. It says nothing about ChatGPT, Gemini, Perplexity or Google AI Overviews. Partner integrations are not named.
• Sampled and still changing. Bing says the grounding-query data is a sample and may be refined, so treat small movements as noise.
• Short history. Momentic reports the lookback starts on November 1, 2025, and retention is unclear, so export what you need.
• Hard to analyse in the tool. Comparing periods or filtering to a subfolder mostly means exporting to a spreadsheet.

One more caution. Some write-ups describe the report as showing AI impressions and clicks. Bing’s own announcement and the more detailed coverage do not, so check your own report before repeating that claim to leadership.

Why B2B teams should care anyway

Bing is a small share of search, but the report is useful beyond Bing for three reasons.

• It is the only first-party AI citation data available. Third-party AI-visibility tools sample answers from outside. This is the engine reporting on itself.
• Grounding queries show how an AI breaks down a topic. They reveal which sub-topics an assistant treats as relevant to your category, which is hard to see anywhere else. Momentic suggests reading them as topics, not keywords.
• Citations concentrate. In one B2B industrial example Momentic described, 86% of grounding queries sat in the core product category, and the top explainer page earned about ten times the citations of the second-best page. That is one case, not a benchmark, but it is a useful pattern to look for in your own data.

A 30-minute routine for B2B teams

1. Verify your site in Bing Webmaster Tools, or import it from Search Console.
2. Export the report and save it with the month in the file name. This builds the baseline the tool does not keep for you.
3. Bucket the grounding queries into five or six topics. Check whether the topics AI associates with you match the ones you sell on.
4. Rank your cited pages. Note what the top ones share, such as structure, depth, freshness and evidence.
5. List important pages with zero citations. Bing suggests finding pages that are indexed but rarely cited and improving them with clear headings, tables, FAQs and sourced claims.
6. Notify Bing of changes using IndexNow when you update key pages.
7. Repeat monthly and report the trend, not a single month.

What not to do

• Don’t call it your AI traffic. Citations are not clicks. Pair the report with branded search, direct visits and self-reported attribution on forms.
• Don’t generalise to every AI engine. It describes Microsoft’s surfaces only.
• Don’t chase grounding queries as keywords. They are the AI’s retrieval phrases, not search demand.
• Don’t over-read a one-week change. The data is sampled and may be revised.

BuzzIQ’s take

The interesting part of this report is not Bing. It is that a search engine has started publishing citation data, which makes AI visibility something you can measure from the source instead of estimating from a vendor sample. We expect others to follow, and teams that already have a monthly export habit will have a head start.

Our strategic takeaways:

• Start the baseline now. Free, first-party and imperfect beats waiting for a perfect tool.
• Judge content by citations and pipeline together. Citations show you are being used as a source. Pipeline shows whether it mattered.
• Fix the pages that should be cited and are not. That list is usually the most actionable output of the whole report.

Where this applies at BuzzIQ. Our AEO and GEO services include baselining how AI engines cite and describe you, and our B2B SEO and AI search work turns that into a content plan. If you have not opened this report yet, start there.

Sources

• Bing Webmaster Blog: Introducing AI Performance in Bing Webmaster Tools Public Preview
• Practical Ecommerce: Bing’s improved AI Performance section
• Momentic Marketing: Initial thoughts on AI Performance reports in Bing WMT
• Locomotive: AI Performance in Bing Webmaster Tools

The report is a public preview and its features are changing, so check your own account before quoting numbers. Case figures are from the cited third-party write-ups, not BuzzIQ data.