Bing Webmaster Tools quietly became one of the few free, first-party windows into how an AI engine uses your content. As of June 2026 it reports which AI "grounding queries" cited your site, what share of citations you captured, the intent behind those queries, and the topics Microsoft's systems associate with you. That's real evidence most teams aren't looking at — but it answers a narrower question than it first appears.
This is part of the measuring AI visibility series; it assumes you already know why citations replaced rankings and focuses on one concrete data source.
What does Bing Webmaster Tools now report about AI?
Microsoft launched an AI Performance report in public preview in February 2026, surfacing the grounding queries (the searches an AI surface runs to ground its answer) where your site was cited. In June 2026 it added four capabilities, rolling out in preview globally:
| Feature | What it shows |
|---|---|
| Intents | Grounding queries classified by intent — Informational, Commercial, Navigational, Research, Local, "Learn and Solve," Creation, and more |
| Topics | A thematic view that groups grounding queries into subject areas where you're gaining (or missing) AI visibility |
| Citation Share | The percentage of AI citations your site captured for a grounding query, out of all citations shown across all sites for it |
| Compare | An overlay of a previous period (e.g. the prior 30 days, or a custom range) to see how citation activity changed |
All figures are attributed to Microsoft's Bing Webmaster blog and corroborating coverage (Search Engine Land, Search Engine Journal) from June 2026. The headline addition is Citation Share — your share of citations on a query, not just a raw count — which is the closest a free tool gets to an answer-level share of voice.
Why does this data matter when it's "only" Bing?
Because Bing is bigger than its consumer search share suggests, and its index feeds AI you care about. Microsoft's web index powers Copilot across Windows and Microsoft 365, and — as reported through 2026 — is among the web-results providers ChatGPT Search can draw on. So a grounding query that cited you in Bing's data is direct evidence your content is retrievable and quotable in the kind of pipeline several assistants share. It's a proxy, not a guarantee — hedge it — but a free proxy backed by first-party data beats inference.
The most useful, least-obvious move is to read Intents and Topics as prompt research. Most teams build a tracking list by brainstorming questions; AI answers are instead assembled from retrieval paths and query fan-out. When Bing groups your citations by intent and topic, it's handing you first-party evidence of the clusters AI already associates with your content — a better starting point than a guess. That feeds directly into how to choose which prompts to track and the wider question of how to measure your AI visibility.
What can't it tell you?
This is where most read it wrong. Citation Share tells you how you did on queries where you already appeared; it is silent on the more valuable question — where you should appear but don't.
Bing Webmaster Tools answers "how did we do?" — not "where should we appear but don't?" It's a free first-party signal, not a cross-engine scoreboard, and the gap is the whole game.
Concretely, the report does not show:
- The exact prompts users typed — you see grounding queries and intents, not the raw conversation.
- The full AI answer your content appeared in, or whether you were recommended versus merely cited (a distinction that matters — see cited isn't recommended).
- Visibility across other engines — Google AI Overviews, Gemini, Perplexity, and Claude are all outside it.
- The "should" list — the topics and queries where you're absent entirely, which by definition can't show up in a report of where you were cited.
So treat it as one strong input, not the system of record. It's the AI-era counterpart to reading your server logs to see which AI bots crawl your site: authoritative for what it covers, blind beyond it.
How to use it in a real workflow
Fold the Bing data into the reporting stack rather than treating it as the whole picture:
- Verify your site in Bing Webmaster Tools (free) and open the AI Performance report.
- Read Topics for authority and gaps — note where you're gaining citation share and which adjacent topics are thin.
- Turn Intents into prompts — use the commercial and "learn and solve" clusters to seed realistic prompts, then validate them against actual customer language.
- Use Compare for trend, not snapshots — citation activity is noisy; a 30-versus-30 overlay shows direction.
- Cross-check across engines — Bing covers Microsoft's surfaces; pair it with the broader five-metric reporting stack so a single engine doesn't define your view.
Bing Webmaster Tools is the rare free, first-party AI signal — worth wiring in today. But "how did we do on Bing?" is one engine's slice of a question that spans all of them: where are we recommended, across every AI surface, over time? Closing that gap — turning per-engine signals into one cross-engine scoreboard — is exactly what Buffy Intel does. Questions: [email protected].
