Bing's Topics Portfolio: what each number means and how to choose your next article
Bing now sizes AI demand by topic and lists the questions where other sites are cited and yours is not. On Index Lab's own data, 34 of the 50 questions we missed in our main topic were hiring questions.
Guilherme Hortinha · 20 October 2026
Bing's Topics Portfolio groups the questions its AI answers with your pages into topics, sizes each one, and lists the questions where other sites are cited and yours is not. That list is a queue of articles worth writing.
What Topics Portfolio is
Bing added Topics Portfolio to Bing Webmaster Tools on 13 October 2026, as a beta tab beside Citations & Pages in AI Performance, according to Bing's launch post. AI Performance, in public preview since February 2026, shows when your pages are cited in Microsoft Copilot, AI summaries in Bing and select partner integrations, plus the grounding queries behind those citations: the phrases the AI uses to fetch what it cites.
Topics Portfolio clusters those queries into themes. It is built from your own site's data, and the tab appears only once a site has enough of it.
What each number means
Each topic row carries three numbers, and opening a topic adds your own citations, your grounding queries and the queries you are missing.
Number | What Bing counts | How to read it |
|---|---|---|
Topic Size | Approximate citations across all sites in the topic | Demand, measured in citations |
Sites Cited | Approximate number of sites cited in the topic | How crowded it is, read beside Topic Size |
Queries | Approximate distinct grounding queries | How many different questions the topic holds |
Opportunity queries | The strongest queries where other sites are cited and yours is not, each with a citation range | Your shortlist of pages to write |
Window | One, two or three months | Keep it the same between exports you compare |
What the beta does not tell you
Bing calls the numbers directional signals for planning. Topic labels are generated by AI and can be wrong. Topics with too little activity are hidden. Branded queries can appear as opportunities. Bing's launch post does not say how citation ranges are calculated, so we read them as relative size only.
See which brands ChatGPT names when your buyers ask their questions.
The CSV exports do not record which window you chose, so put it in the file name, and the launch post mentions no API. Every topic and opportunity carries a thumbs up and a thumbs down, which is how Bing asks to hear that a row does not fit.
What it showed for indexlab.ai
Index Lab's own data, July to September 2026: Bing's AI cited indexlab.ai 49,433 times across 471 grounding queries, and 92% of those citations, 45,607, sat in one topic, SEO & Search Optimization.
They came from research and tool questions: which SEO platforms track AI SEO projects, how named AI visibility tools compare, how Gemini changes brand discovery. Queries that mention an agency added up to 11 queries and 384 citations, under 1% of the total, and most asked about tools agencies use. Almost nobody reached us while trying to hire.
Of the 50 opportunity queries for SEO & Search Optimization, 34 ask which agency or service to hire, 6 ask for tools and 10 ask how to do something. Six sit in the 10K to 50K citation range. So our next pages answer hiring questions.
Why the list is only as specific as your non-branded citations
Opportunities come from the topics where you are already cited, so the list is only as precise as your citations. We are cited across one niche, and the list came back full of that niche's buyer questions.
A site that Bing cites only for its own name gets broad suggestions: questions from other countries and languages, with no link to its market. The list narrows as the site earns citations on questions that do not contain its name.
How to turn the list into a writing plan
Bing advises starting in the topics where your citations already concentrate. Export three months and note the window. Sort that topic's opportunities by what the asker wants: to hire, to buy a tool, or to learn how. Thumbs down the ones outside your market. Merge near-duplicates; six of our 50 are variants of "top ai seo agencies". Then write one page per group.
In July 2026 the keyword data we pulled showed no search volume for "aeo agency" or "ai seo services", while Bing lists questions built on those words at 1K to 5K citations. Re-export monthly and track your citations against Topic Size, or ask Index Lab to measure the same gap across the other AI engines.
Frequently asked questions
Is Topic Size the number of searches?+
No. Topic Size counts approximate citations across every site in the topic over the window you chose. It measures how often AI answers in that topic cite sources.
Can I pull Topics Portfolio data through an API?+
Bing's launch post describes no API for it. The data comes out as CSV downloads, one for the portfolio and one per topic you open.
Why does my opportunity list look irrelevant?+
Often because Bing cites your site for few questions beyond your own name. It can also be a labelling miss, since AI classifies the topics. Thumbs down what does not fit, publish pages that earn non-branded citations, and compare next month's export.
How is this different from Google Search Console's generative AI report?+
Google's generative AI report shows impressions with no query data: you learn that you appeared, never on which question. Bing shows the grounding queries, and Topics Portfolio adds the demand around them.
Find out if your site is technically ready to be quoted by AI.
[ Keep exploring ]
Terms in this article
Co-founder & Head of AI Innovation, Index Lab
Guilherme co-founded Index Lab, an AEO/GEO agency that makes brands the answer AI gives across ChatGPT, Gemini and Perplexity, taking clients from zero AI visibility to top recommendations.
[ Related reading ]
ChatGPT Ads for Local Services: 90 Million Weekly Requests and a Gap Worth Knowing
OpenAI's exploratory estimate puts requests for nearby businesses and service providers at over 90 million ChatGPT messages a week. Why local demand on ChatGPT is a fraction of what it is on Google, and when that gap works in a local business's favour.
Read →AI Search EcosystemChatGPT Ads for B2B Software: 490 Million Weekly Messages and What a Click Costs
OpenAI counts over 490 million weekly ChatGPT messages categorised as B2B or B2B-adjacent, and a further 90 million on small-business operations. How the demand reads, what the published B2B cases show, and what a click costs on average across the campaigns we manage.
Read →AI Search EcosystemChatGPT Ads for Healthcare and Clinics: 2.4 Billion Weekly Messages, With a Policy Gate
OpenAI counts over 2.4 billion weekly ChatGPT messages about care, fitness, nutrition, test results and symptoms, up 31% since January. Health advertisers have been admitted since August 2026. What the number covers, and what the review requires.
Read →See where your brand stands in AI.
Free visibility check →