ObsurfableFor two years, publishers have asked the same question about AI search: Is my content actually being cited? Google Search Console added AI Overview data, but no major platform offered dedicated citation reporting — until Microsoft shipped it.
In February 2026, Microsoft introduced the AI Performance Report in Bing Webmaster Tools. In June 2026, it expanded with four preview capabilities: Intents, Topics, Citation Share, and Compare. For the first time, verified site owners can see which pages are cited in AI-generated answers, which grounding queries triggered retrieval, and how citation activity trends over time — across Microsoft Copilot, Bing AI summaries, and select partner integrations (including ChatGPT, which retrieves through Bing).
This is the most significant measurement infrastructure release in AI visibility since the category emerged. It is also incomplete — and understanding both halves matters.
What the AI Performance Report shows
The dashboard provides five core metrics, refreshed daily with a short processing delay:
| Metric | What it measures |
|---|---|
| Total citations | How many times your content was referenced as a source in AI-generated answers |
| Average cited pages | Daily average of unique URLs from your site cited across AI experiences |
| Grounding queries | Key phrases the AI used when retrieving your content (not necessarily the user's search query) |
| Page-level citation activity | Citation counts per URL — which specific pages are referenced most |
| Visibility trends | Timeline showing how citation volume changes over time |
Grounding queries: the most actionable metric
Grounding queries deserve special attention. They are not what users typed. They are the phrases the AI system used internally to find and pull your content into an answer.
Microsoft describes them as "grouped representations from your site that were retrieved and cited in AI-generated answers." A single grounding query can map to multiple pages. A single page can appear under multiple grounding queries.
This is the closest any platform has come to showing which questions your content answers in the AI layer — not which keywords you rank for in traditional search, but which retrieval paths lead to your pages being cited.
Example: your pricing page might appear under grounding queries like "CRM pricing comparison 2026" and "affordable project management tools" — revealing the AI-intent paths driving citation, independent of your organic keyword strategy.
Citation Share: relative visibility within a query
The June 2026 expansion added Citation Share — calculated as the percentage of citations attributed to your site out of all citations shown across all sites for the same grounding query.
Total citation counts tell you whether you appear. Citation Share tells you how much of the citation space you occupy for a specific query. A page cited once in an answer where ten sources appear has a different strategic value than a page cited once where two sources appear.
Citation Share is the first platform-native share-of-voice metric for AI citations.
Intents and Topics: query context beyond keywords
The preview capabilities classify grounding queries by intent (informational, commercial, navigational) and topic clusters. This moves reporting from "you were cited X times" to "you were cited X times for commercial-intent queries about [topic]" — enabling content strategy decisions grounded in AI retrieval patterns.
Compare: competitive context
The Compare view lets publishers benchmark citation activity across time periods or query groups — spotting whether a content update, model change, or competitive shift moved citation share up or down.
What the report covers — and what it does not
Coverage
The report tracks citations across:
- Microsoft Copilot
- AI-generated summaries in Bing
- Select partner AI integrations (Microsoft has not published a complete partner list, but ChatGPT's Bing retrieval is the most significant inclusion)
Because ChatGPT retrieves primarily through Bing's index, Bing's AI Performance data is a proxy for ChatGPT citation activity — not a complete picture, but the closest platform-native measurement available.
Limitations
Microsoft is explicit about what the report does not measure:
No click-through data. You can see that your content was cited. You cannot see whether users clicked through from the AI answer to your site. Fabrice Canel from Microsoft confirmed on X: "With this preview, the data is not yet available via the API."
Aggregated, not exhaustive. "AI Performance data is aggregated and summarized to provide a representative view of citation activity. The data shown is not a complete log of every instance where content from your site was referenced."
Citation ≠ placement or prominence. Total citations count source references without indicating where in the answer your content appeared, how prominently it was featured, or whether your brand was named in the answer text. Ghost citations — cited but not named — are invisible in this data.
No brand mention tracking. The report tracks URL-level citation activity. It does not report whether your brand name appeared in the answer text alongside the citation.
No cross-engine comparison. Bing's report covers Microsoft's ecosystem. It does not show Claude (Brave), Perplexity (proprietary), or Google AI (Google index) citation activity for the same queries.
No API access yet. Data is dashboard-only. API availability is on Microsoft's backlog.
Why this matters for publishers
1. Citation is now a reportable metric — for one ecosystem
Before February 2026, AI citation tracking required third-party tools running prompts manually or at scale. Bing Webmaster Tools made it a native platform metric — at least for the Microsoft/Bing/ChatGPT retrieval stack.
This legitimizes citation as a KPI alongside clicks, impressions, and rankings. Publishers can now include "AI citations" in monthly reporting with platform-sourced data, not just vendor estimates.
2. Grounding queries reveal AI-intent paths
Traditional keyword research shows what users search for. Grounding queries show what the AI retrieval system searches for when assembling an answer. These are different.
A page ranking for "best CRM software" might be cited via grounding queries like "CRM comparison for small teams" or "Salesforce alternatives 2026" — revealing sub-intent paths that keyword tools miss. This is actionable for content gap analysis.
3. Page-level data enables content-level optimization
Knowing which specific URLs are cited — and which are not — lets teams optimize at the page level. A comprehensive guide that ranks well organically but never appears in AI citations needs different treatment than a mid-ranking page that is cited frequently. Page-level citation data makes that distinction visible.
4. ChatGPT visibility becomes partially measurable
Because ChatGPT retrieves through Bing, improving Bing AI Performance metrics likely correlates with improved ChatGPT citation rates. Seer Interactive found 87% overlap between ChatGPT Search citations and Bing top results. Bing Webmaster Tools is the closest thing to a ChatGPT citation dashboard that exists today.
How to use the report: a practical workflow
Step 1: Verify your site in Bing Webmaster Tools
If you have only verified in Google Search Console, add Bing verification. Submit your sitemap. Enable IndexNow for faster indexing. Allow OAI-SearchBot in robots.txt.
Step 2: Baseline your citation metrics
Record total citations, average cited pages, and top grounding queries for the past 30 days. This is your starting point — not a benchmark, since historical data availability is limited.
Step 3: Map grounding queries to content gaps
For each grounding query where you have low or zero Citation Share:
- Do you have a page that answers this query directly?
- Is a competitor cited instead? Which page?
- Does the query represent a buyer-intent path you have not covered?
Feed gaps into your content calendar with AI citation as the success metric.
Step 4: Cross-reference with Google Search Console
Compare Bing AI Performance grounding queries with Google Search Console query data. Queries appearing in both ecosystems are high-priority. Queries appearing only in Bing AI Performance may represent ChatGPT-specific retrieval paths.
Step 5: Track Citation Share trends after content updates
When you publish or refresh a page, monitor Citation Share for related grounding queries over the following 2–4 weeks. With median citation lifespan at 11–15 days (Writesonic), expect movement within that window — not months.
Step 6: Supplement with cross-engine monitoring
Bing's report covers one retrieval backend. Run the same grounding queries through Claude (Brave), Perplexity (proprietary), and Google AI (Google index) to see where Bing citation gains do not carry over.
What publishers still need beyond Bing
Bing's AI Performance Report is the best platform-native citation tool available in July 2026. It is also one ecosystem out of four.
| Need | Bing AI Performance | Still requires |
|---|---|---|
| ChatGPT/Bing citations | Yes | — |
| Claude citations | No | Parallel monitoring (Brave backend) |
| Perplexity citations | No | Parallel monitoring (proprietary index) |
| Google AI citations | No | Google Search Console (partial) |
| Brand mention tracking | No | Answer-text analysis |
| Cross-engine comparison | No | Multi-platform tool |
| Citation persistence/volatility | Partial (trends) | Repeated prompt runs |
| Competitor citation share | Partial (Compare) | Prompt-level competitive tracking |
Google Search Console reports AI Overview appearance data but lacks grounding queries, Citation Share, and page-level citation counts in the same depth. No platform offers Claude or Perplexity citation reporting natively.
The gap between "Bing shows my citations" and "I know my full AI visibility" is where multi-platform monitoring tools operate.
How Obsurfable fits
Obsurfable complements Bing's AI Performance Report by monitoring the engines Bing does not cover — Claude, Perplexity, Google AI Mode, and Gemini — on the same buyer prompts. When Bing's grounding queries show citation activity for "best CRM for startups," Obsurfable shows whether the same query produces citations on ChatGPT, Claude, and Perplexity too.
The Visibility Director cross-references platform-native data with cross-engine gaps and drafts content for the backends where visibility is missing.
For ChatGPT's Bing dependency, see Why Each AI Engine Reads a Different Search Index. For measurement beyond citations, see How to Measure AI Visibility: KPIs Beyond CTR.
FAQ
Does Bing AI Performance include ChatGPT citations?
Microsoft includes "select partner AI integrations" without naming all partners. ChatGPT retrieves primarily through Bing's index (87% citation overlap per Seer Interactive), so Bing AI Performance is a strong proxy — but not a guaranteed 1:1 mapping.
Can I get this data via API?
Not yet. Microsoft confirmed API access is on the backlog. Dashboard-only for now.
How does this compare to Google Search Console AI data?
Google Search Console reports AI Overview appearance impressions and clicks. Bing's report goes deeper on citation-specific metrics: grounding queries, page-level citations, and Citation Share. Neither covers non-native engines.
Should I prioritize Bing optimization now?
If your buyers use ChatGPT or Copilot, yes. Index in Bing, allow OAI-SearchBot, use IndexNow, and monitor AI Performance. But do not treat Bing metrics as full AI visibility — Claude and Perplexity require separate strategies.
Bottom line
Bing Webmaster Tools' AI Performance Report is the first major platform to answer "which pages does AI actually cite?" — with grounding queries, page-level data, and Citation Share. Use it for the Microsoft/Bing/ChatGPT stack. Supplement it with cross-engine monitoring for Claude, Perplexity, and Google AI — because one platform's citation dashboard cannot show you what the other three backends are doing with the same buyer questions.