Bing's AI Performance report now shows the queries where rivals get cited and you do not. Read it as a sample.

On 13 October 2026 Microsoft added a tab called Topics Portfolio to the AI Performance report in Bing Webmaster Tools.

The post is by Krishna Madhavan, Saral Nigam and Vivek Jain of Microsoft Bing (Bing Webmaster Blog, Introducing Topics Portfolio (Beta) in AI Performance, 13 October 2026, read on 6 November 2026).

The headline feature is a list of queries where other sites get cited in AI answers and yours does not.

Google's generative AI report in Search Console has nothing like it.

But here is the complication. Bing says the tab only appears when there is "sufficient data for your site," and that smaller sites "may see a shorter list or may not even see this feature." A single-location business, the kind NovaTechRay mostly works with, can open the report and find an empty tab.

So the most useful competitor view in AI search may not be available to the businesses that need it most.

Step 1: What the AI Performance report counts

The report itself is older than the tab.

Bing opened it as a public preview on 10 February 2026 (Bing Webmaster Blog, Introducing AI Performance in Bing Webmaster Tools Public Preview, read on 6 November 2026).

That post names three places it counts: Microsoft Copilot, AI-generated summaries in Bing, and "select partner integrations." It does not list the partners.

It shows five things. Total citations, average cited pages per day, grounding queries, citation counts per URL, and a timeline.

Grounding queries are the phrases the AI used when it went looking for your content. That is the part no Google report gives you, and it is why we called it the only place you can read the sub-questions directly in why every AI visibility tool gives you a different number.

There is no click column. Barry Schwartz said the same of the new tab on 13 October: "No click data yet, sorry folks..." (Search Engine Roundtable, Topics Portfolio In Bing Webmaster Tools AI Performance Reports, read on 17 October 2026).

Step 2: What Topics Portfolio adds

The new tab sits next to Citations & Pages and groups your grounding queries into themes.

Bing's examples of themes are football, taxes and AI chatbots. The grouping is AI-generated, and Bing does not describe the method.

Each topic row carries three numbers:

1. Topic Size, the approximate total citations across all sites for that topic in the period.
2. Sites Cited, the approximate number of sites that appear for it.
3. Queries, the approximate number of distinct grounding queries inside it.

Click into a topic and you get your own citations in it, your grounding queries, and Opportunity Queries: the ones where other sites appear and yours does not.

A portfolio-wide list called Top Opportunity Queries sits under the table. Windows are one, two or three months.

Matt G. Southern of Search Engine Journal points out what changed from earlier versions: the older Topics labels only grouped queries where you were already cited (Search Engine Journal, Bing Webmaster Tools Sorts AI Search Queries Into Topics, 13 October 2026, read on 17 October 2026). The new part is the queries you are missing.

Step 3: Read every number as a sample

Bing is unusually honest about its own limits here, and the honesty is the most useful part of the post.

The February post calls grounding queries "a sample of overall citation activity." It also says citation counts do not indicate ranking, authority, or a page's role inside an answer.

The October post goes further. Its numbers are "directional signals for planning," not "a precise ledger." Topic labels can be wrong. Opportunity volumes are shown as ranges on purpose.

So a rising citation count does not tell you that you moved up, or that anyone read your name. It tells you Bing's sample saw your page used more often.

We would read Topics the way we read the Search Console generative AI report, which counts impressions and nothing else (our walkthrough of that report). Useful for direction. Useless as a scoreboard.

Webmaster consoles only report what the engine chooses to tell you. When the group checked the Manual actions report across all six VALORAE sites on 9 October, a clean result ruled out a human penalty and nothing more. Same discipline here.

Step 4: Opportunity Queries are not all opportunities

Ignore half the list.

Bing's own example shows why. A made-up shoe shop sees a query for a rival's branded running shoe listed as an opportunity. Bing flags that branded queries "can appear as opportunity queries."

For a local business that will be common. Queries naming a competitor, queries for a service you do not offer, queries for an area you do not serve.

None of those are gaps. Writing a page to chase them is how sites end up with the thin, scaled content that Google's spam policies target, a risk we covered in the September 2026 spam update post.

The useful opportunities are the boring ones. A question about your exact service, where a directory or a rival's page is cited because yours never answered it in plain text.

Every row has a thumbs up and thumbs down. Use them. Bing says the feedback goes to the team working on classification and opportunity selection, and a beta tab is shaped by whoever bothers to vote.

Step 5: If your tab is empty

Bing has not published what "sufficient data" means.

Southern also notes that, as of his writing, Bing's AI Performance help page did not yet cover Topics Portfolio, so the announcement is the main documentation. Bing gives no date for the tab leaving beta, and on 6 November 2026 its Webmaster Blog listed nothing newer than the 13 October post.

An empty tab does not mean Copilot ignores you. It means Bing does not have enough of a sample to show you topics.

The Citations & Pages view can still have data when Topics does not. Start there. Write down which of your URLs are cited and which grounding queries appear, with the date.

Then run the manual method from how to measure whether AI assistants are recommending you: a fixed prompt set, several runs each, recorded in a sheet. That post was written before we gave Bing's report its own column, and a dated note on it now points here.

Step 6: Where this fits in a NovaTechRay test sheet

Copilot gets its own column, fed by Bing's report rather than by guessing.

Bing's report says nothing about ChatGPT, Perplexity, Claude or Google. Bing does not list its partner integrations, so we do not assume any of those are inside the count.

That leaves three sources of truth across the major surfaces, none complete. Search Console gives Google impressions. Bing gives Microsoft citations and a sample of grounding queries. Everything else is your own sampling, run by hand or by a tracker.

What changed on 13 October is that one of the three now shows competitors. When a NovaTechRay client has enough Bing data for the Topics tab, we will read its opportunity list against the prompt set and add any real gap as a new prompt, not as a new page.

We have not yet seen how stable the topic labels are from month to month. If Bing publishes that, or the thresholds, we will update this post with the date.

The short version

• Verify the site in Bing Webmaster Tools and open AI Performance.
• Read Citations & Pages first, because Topics may be empty for a small site.
• Treat every count as a sample, as Bing itself says it is.
• Cross out branded and out-of-scope opportunity queries before acting.
• Answer the real gaps in plain text on the page that already owns the topic.
• Keep Copilot in its own column, apart from Google, ChatGPT and Perplexity.
• Expect no click data from this report yet.

We keep a column like this for every client at novatechray.com.