Bing Webmaster Tools: the AI view that reveals your missed citations
Summarize this article with AI
What does Topics Portfolio change in Bing Webmaster Tools?
I look at Bing Webmaster Tools accounts for my clients. Most teams open them once a quarter. For a sitemap. For an indexing error. Then they close the tab.
That habit needs to change. On October 13, 2026, Microsoft announced a beta feature called Topics Portfolio. It sits in the AI Performance report, in a new tab to the right of « Citations & Pages ». According to Search Engine Journal (article by Matt G. Southern), this view does two things.
- It groups the queries for which your site is cited in AI answers into broader topics.
- It lists the queries within those topics where other sites are cited and yours isn’t.
The second point matters more than the first.
Bing calls these queries grounding queries. They are the search terms Bing’s AI uses to locate content, then cite it in its answers. That covers Copilot, Bing’s AI summaries and some partner integrations. In other words: these are not the words your customers type. They are the words the machine uses to go and find a source.
Until now, the AI Performance report told you where you were cited. It entered public preview in February 2026. In June, Microsoft added Topics, Intents, Citation Share and Compare. But those groupings only covered queries where you already appeared. You saw your own garden. Never the neighboring field.
The change fits in one sentence. You finally see what you’re missing.
Key takeaway: Topics Portfolio only shows up if Bing has enough data on your site. Topics with too little presence or activity in the ecosystem are hidden. A small site may therefore see few items, or even none. That’s not a configuration error. It’s Bing’s threshold.
Why am I telling you about it now? Because this feature is free. It works today. And you start now, without waiting for Google to open its citation data. Your most attentive competitor will open it this week. Is your Bing account already verified and populated?
How do you read Topic Size, Sites Cited and Queries?
The first screen of Topics Portfolio shows the topics where Bing estimates your site has a « meaningful » presence. Each row carries three metrics. They don’t say the same thing. Mixing them up leads to bad decisions.
| Metric | What Bing measures | What I do with it |
|---|---|---|
| Topic Size | Estimate of the total number of citations, all sites combined, over the chosen period | I use it as a demand barometer |
| Sites Cited | Approximate number of sites that appear on the topic’s queries | I use it as a competition barometer |
| Queries | Approximate number of distinct grounding queries in the topic | I use it as a breadth barometer |
Microsoft describes Topic Size as a way to gauge search demand flowing through its AI. That’s exactly how I read it. A large topic with few sites cited is open ground. A narrow topic crowded with sites is closed ground.
The cross-reading of the three columns tells you everything.
- Large topic, few sites cited, many queries: demand exists, supply is thin, the zone is wide. High priority.
- Large topic, many sites cited, few queries: a tight market on a handful of questions. Target one precise query, not the whole topic.
- Small topic, few sites: a niche. Useful if it fits your catalog, worthless otherwise.
The issue is never raw volume. It’s the ratio between demand and competition. I’ve practiced this reading for years on the semantic cocoons I build: a theme only has value through the gap between what people search for and what the market covers.
Another detail. You choose the window: one, two or three months. I always start with three months. Why? A short window exaggerates accidents. A news item, a promo, a seasonal spike. Three months smooth the noise. Only then do I drop to one month to check a trend.
Volume figures are given as ranges, not exact values. That keeps the precision honest. You compare orders of magnitude, not decimals. When you open your topics, which one shows the strongest demand and the weakest competition?
What does the Top Opportunity Queries list contain?
Below your topics table, Bing displays a list called Top Opportunity Queries. Across all your topics, it gathers the queries where other sites are cited and yours isn’t. Citation volumes are given as ranges.
This is the heart of the tool.
In my line of work, we operated blind on AI search for a long time. We could test prompts by hand. We could count mentions on a sample. But nobody could say: « on this precise query, Bing cites other sites and not yours, with this order of magnitude of citations. » Today, Bing tells you.
One click on a topic opens a detail page. There you find the topic’s size, your citations within that topic and its number of queries. Two tables complete it.
- The first lists the topic’s queries where your site already appears.
- The second lists the opportunity queries, again with citation ranges.
I read them side by side. The first shows what Bing understands about your expertise. The second shows where it pulls from elsewhere what you don’t offer yet. Between the two, you find your next page.
A field observation, with no figure to sell you: in the audits I run, visibility gaps don’t come from an overall content deficit. They are neighboring questions to existing pages, never addressed. The site talks about trail running shoes, but no page answers the question that comes before the purchase. The content exists. The answer is missing.
Citation Share, launched in June, adds a nuance. It measures your share for a grounding query. According to Bing’s help page, it never reveals competitor domain names. You see your relative position without access to other sites’ data. Topics Portfolio works the same way: it signals citations elsewhere without showing who holds them. It’s up to you to observe the answers in Copilot to grasp the cited format.
What is the first query on your list that you had never thought of covering?
How do you turn a missed query into a cited page?
An opportunity list produces nothing on its own. I apply this five-step method. It draws on my practice with cocoons and on Bing’s recommendations.
Step 1: sort opportunities by subject. Don’t work query by query. Open each subject and group the high-value queries that look alike. Three neighboring queries often call for a single well-built page, not three thin pages.
Step 2: check what you already have. Before creating, search your site. An existing page that half-covers the question is worth more than a new page competing with your own. Internal cannibalization is costly. It blurs the signals.
Step 3: read the matching AI answer. Type the query into Copilot. Look at the cited sources. Observe the format: short definition, comparison table, step list, direct answer at the top of the page? The cited format tells you what the machine knows how to extract.
Step 4: write to be extracted. A direct answer in the first lines. Subheadings that echo the query’s wording. Verifiable data. A structure that lets a system grab a piece of information without reading the whole text.
Step 5: connect to the architecture. Each new page hooks onto a parent page and its siblings through explicit internal links. An isolated page is a fragile page.
The diagnosis: it’s an architecture problem, not a content problem. When a query slips away from you, ask the right question: where in my cocoon is this answer missing?
I often see the opposite among e-merchants. They launch a series of blog posts to fill a gap, disconnected from the category pages. The articles get published. Nothing moves up. The production happened in the wrong place.
I make no numerical promises here: Bing guarantees nothing, and neither do I. What I can do is give you a repeatable method. You handle one subject at a time. You re-measure after a cycle of one to three months. You keep what works.
How many of your existing pages could absorb a high-value query with just one extra section?
What are the beta’s limits, and why do they work in your favor?
Microsoft is clear about it itself: Topics Portfolio is a beta. According to Search Engine Journal, three limits deserve your attention.
1. Topics are AI-generated. A query may be filed under a label you wouldn’t have chosen. Your own category logic is not Bing’s. This breaks nothing. It simply asks you to read the groupings with a critical eye.
2. Brand queries can show up as opportunities. Microsoft gave an example: a shoe site saw a competitor’s brand search surface in the « Running Shoes » topic as an opportunity. Obviously, you won’t create a page to be cited on your competitor’s name. That sorting is yours to do.
3. Figures are ranges. Microsoft recommends treating them as directional signals for planning. I agree with that stance.
And here’s the counter-intuitive point I stand by: these limits are good news.
Why? Because a perfect tool would have been used by everyone in the same way, in the same order, with the same conclusions. Imperfect data rewards judgment. Whoever can tell an opportunity from a brand query gains ground on whoever copies the list.
In my view, the trap lies elsewhere. It’s the temptation to cover everything. The list is long? You’ll be tempted to produce ten pages at once. Bad reflex. A visibility gap is not an order. It’s a hypothesis to test. You pick three, build them well, measure.
Keep the threshold constraint in mind too. If your site is small, Topics Portfolio may stay empty or close to it. In that case, don’t force it. Keep using the other AI Performance views (Citations & Pages, Citation Share, Compare) and keep building your baseline presence. The view will open once the data is sufficient.
If you had to sacrifice 80% of the list and keep only what fits your business, what would remain?
Why wait for Google to open its own data when you can act today?
In meetings, I often hear: « We’re waiting to see what Google does. » I understand the instinct. Google is the reference. But it’s the wrong urgency.
Bing now puts at your disposal a report that shows your citations in AI answers, your topics, your citation share, a comparison, and now a list of opportunities. The channels involved, according to the source, are Copilot, Bing’s AI summaries and partner integrations. I’ll say nothing more about the reach of these channels, because the source doesn’t give its scale and I won’t invent it.
The data exists, it’s free, and it comes straight from the engine. In a field where what we read rests mostly on manual tests and estimates, that’s rare. Waiting means choosing to stay in estimation.
There’s a second, more structural argument. The work Topics Portfolio pushes you to do is transferable. Spotting uncovered neighboring questions. Structuring an extractable answer. Connecting each page to a cocoon. None of this is specific to Bing. It’s good semantic architecture. What you build to fill a gap at Microsoft also serves your pages in other engines, because the same principles of clarity and structure apply everywhere. In my view, that transfer is the benefit.
Third argument: the head start. The opportunities displayed today are there because other sites haven’t answered yet. Once they open the view, the list will shrink. Whoever acts first takes the spot. Whoever acts last fights for it.
I know what this implies: adding a routine to an already full calendar. So I suggest a minimal format. Half an hour a month. Open the view, pick out three opportunities, decide on one action. One decision, with no extra dashboard and no extra report.
Systems that run without you: that’s my core principle. A documented thirty-minute monthly routine that a team member can run in your place beats an exceptional audit that sleeps in a folder.
When was the last time you made an SEO decision based on data supplied directly by an engine?
What action plan can you follow starting this week?
Let’s get practical. Here’s the plan I’d give an e-commerce manager who calls me tomorrow.
Day 1: check access. Log in to Bing Webmaster Tools. Open AI Performance. Look for the new tab to the right of Citations & Pages. If it doesn’t appear, either the beta hasn’t rolled out to your account yet, or your data volume is insufficient. Note it down. Don’t force it.
Day 2: map. Set the window to three months. Export or copy your topics with their Topic Size, their number of Sites Cited and their number of Queries. Sort them by the demand/competition cross-reading described above.
Day 3: sort the opportunities. In Top Opportunity Queries, remove competitor brand queries. Remove those unrelated to your catalog or your expertise. Keep those that naturally extend your existing pages.
Day 4: choose three targets. Not ten. Three. For each, note the closest existing page, the answer format cited in Copilot and the placement in your cocoon.
Week 2: produce. A direct answer at the top. Subsections that echo the query’s wording. Internal links to the parent page and sibling pages. Precise data, sourced or marked as observed.
Following month: measure. Come back to the view. Has the query moved from the opportunities list to the list of queries where you appear? That’s your indicator. It’s simple. It’s readable.
I’ll finish with what I repeat to my clients: keep visibility and traffic separate. A citation in an AI answer is not a session. It creates presence, awareness, a place in the machine’s reasoning. The effects on your sales are measured elsewhere, in your own analytics. Keep both lenses.
I don’t sell you the method. I show you the pages. If you want us to open your Bing account together to see the topics and queries you’re missing, my audit does exactly that. No slide deck. Your own screen.
So, among the queries where your competitors are cited today, which one would you like to see move up into your « Already appears » list next month?
Let’s audit your Bing account live
I walk through your topics and your opportunity queries with you, page by page. You leave with the three gaps to tackle first and their place in your cocoon.
Book a strategic call — 45 minFrequently Asked Questions
What is Topics Portfolio in Bing Webmaster Tools?
It’s a beta view of the AI Performance report. It groups by theme the queries where your site is cited by Bing’s AI, and lists those where other sites are cited and you aren’t.
Why don’t I see the Topics Portfolio tab on my account?
According to Search Engine Journal, the view only appears if Bing has enough data on your site. Small sites see little, or even nothing at all, during the beta.
What are grounding queries at Bing?
They are the search terms Bing uses to find content and cite it in its answers, via Copilot, Bing’s AI summaries and some partner integrations.
Can I see which competitors are cited in my place?
Citation Share keeps competitor domain names hidden. Topics Portfolio shows that citations exist elsewhere, as ranges, while site identities stay private.
Which analysis period should I choose to get the most from the view?
Bing offers one, two or three months. I start with three months to smooth one-off spikes, then switch to one month to confirm a trend before producing a page.
