TLDR;
Bing added an AI Performance preview to Webmaster Tools, showing when content is cited in supported Microsoft and partner experiences. It gives brands direct evidence of exposure and some of the queries behind it. Citations can guide content improvements, but they are not visits or sales.
What happened
Bing introduced AI Performance in Webmaster Tools, reporting how site content is cited across Microsoft Copilot, Bing and selected partner experiences. The preview offered citation activity and grounding-query information, giving site owners evidence of AI exposure beyond traditional organic search rankings.
Why it matters
Bing's preview gives site owners a view of content used in AI responses that conventional ranking reports cannot supply. Grounding-query information can help identify which pages support customer questions. Its practical value depends on coverage and definitions, so enterprise dashboards should preserve the distinction between observed citations and commercial outcomes.
First-party citation reporting gives brands evidence from the provider's own system, which is more useful than relying solely on manually observed answers. It can show which pages and topics receive coverage within the tool's stated scope. That still leaves a substantial distance between being cited and creating a customer or sale. A page may be useful evidence without being the recommended destination, while a citation's commercial importance depends on the question it supports. The enterprise opportunity is to connect observed coverage with content planning: identify where the organisation is contributing useful information and where priority topics lack support. Public preview status and reporting coverage need to remain visible in the interpretation. The tool should improve diagnostic decisions, not become a new scoreboard that implies complete knowledge of the assistant audience.
How your brand can benefit / be affected
Establish a baseline for strategically important pages and inspect the grounding queries where available. Look for factual gaps, weak source pages and unexpectedly cited legacy content.
Compare citation activity with qualified referrals and conversions without treating them as interchangeable. Document preview coverage and any reporting changes. Use the data to choose content investigations and validate outcomes, rather than publishing a single score that claims to represent the entire AI market.
Establish a baseline before making changes. Group cited pages by business purpose and review the associated grounding queries where available. Separate useful commercial questions from incidental topics that generate citations without serving a brand objective. Compare the source page with the approved information and identify whether the citation supports an accurate, helpful answer. If a priority topic is absent, investigate the content's usefulness and accessibility before concluding that the platform is excluding the brand. Reporting is evidence of what the tool observed, not a definitive account of every possible question or answer involving the website.
Assign a small number of improvements to existing content owners, then monitor the relevant pages and topics rather than only the total citation count. Keep changes in tool coverage and product availability annotated so apparent growth is not automatically credited to the content work. Combine the report with referral and commercial data where those signals exist, without assuming a direct causal link. For enterprise reporting, explain the metric in plain language and state its limits next to the chart. A useful programme produces better answers and clearer content priorities. A rising count without an account of relevance can reward low-value exposure and distract from the questions customers actually need resolved.
News date: 10 February 2026. Editorial review: 16 September 2026. Analysis includes subsequent developments where stated.