Bing has given publishers something AI visibility tools have been trying to estimate for months: a first-party view of queries where other sites appear in AI answers and yours does not.
The new Topics Portfolio beta sits inside Bing Webmaster Tools' AI Performance reporting. It groups grounding queries into themes and surfaces Opportunity Queries, including cases where other sites are cited and yours does not appear.
That is useful.
It is also easy to misuse.
Microsoft says the figures are approximate, classification can be wrong, small sites may not have enough data to see the feature, and competitor-branded queries can appear as opportunities.
So the right workflow is not to export every missing query and commission a page.
Validate the gap first.
What Topics Portfolio shows
Topics Portfolio groups AI Performance data by theme rather than making teams inspect queries one at a time.
Microsoft lists fields including:
| Field | What Bing provides |
|---|---|
| Topic | Bing’s thematic grouping |
| Topic Size | Approximate citation volume |
| Sites Cited | Approximate number of cited sites |
| Queries | Approximate query count |
| Top Opportunity Queries | Queries where other sites appear and yours does not |
| Topic Opportunity Queries | Missing-query opportunities within a topic |
Microsoft says the values are shown as ranges, not exact counts.
Reporting windows cover one to three months.
That makes the feature useful for prioritisation, but weak for precise market-share calculations.
Do not turn an approximate range into a percentage with false precision.
Not every opportunity query is your opportunity
One Microsoft caveat deserves special attention: competitor-branded queries can appear in the opportunity list.
Imagine your company sees:
“Acme CRM pricing”
when Acme is a competitor.
Bing may correctly identify that other sites appear and yours does not.
That still may not be a sensible content target.
The intent is brand-specific. Creating a thin page purely to force yourself into that answer could waste editorial resources.
Other opportunities may also be poor fits because they are:
- outside your market;
- outside your product category;
- support queries you do not serve;
- obsolete;
- navigational;
- commercially irrelevant.
Filtering comes before publishing.
Run an AI Citation Gap Audit
Start with 50 to 100 Opportunity Queries rather than the entire report.
For each query, record:
| Field | Audit question |
|---|---|
| Relevance | Does this belong to our market? |
| Intent | Research, comparison, purchase, support or navigation? |
| Current cited sites | Who appears instead? |
| Existing page | Do we already have something relevant? |
| Page quality | Does it answer the task directly? |
| Commercial value | Would visibility matter? |
| Freshness | Is our information current? |
| Manual check | Can we reproduce the gap? |
| Action | Improve, create, consolidate, monitor or ignore |
Manual verification matters because Topics Portfolio is a beta and Microsoft warns that classification errors can occur.
Treat the report as a discovery layer, not ground truth.
Opportunity Query Validation Rate
NEMO would first calculate:
Opportunity Query Validation Rate = relevant opportunity queries after manual review ÷ opportunity queries reviewed
Suppose Bing surfaces 100 opportunities.
After review:
- 18 are competitor-branded;
- 12 are outside your market;
- 10 have mismatched intent;
- 5 are clearly misclassified.
That leaves 55 useful opportunities.
Your Opportunity Query Validation Rate is:
55 ÷ 100 = 55%
The metric tells you how much of Bing’s opportunity list survives human review.
A low rate does not make the feature useless.
It tells you not to automate content production from it.
Fixable Citation Gap Rate
Next, identify opportunities where you already have a relevant page.
NEMO calls this the Fixable Citation Gap Rate:
validated opportunity queries with a relevant existing page ÷ validated opportunity queries reviewed
If 55 opportunities survive validation and 30 already map to a useful URL:
30 ÷ 55 = 54.5%
Those pages should usually be investigated before anything new is written.
Check whether:
- the answer is buried;
- information is outdated;
- the exact comparison is missing;
- headings are vague;
- evidence is difficult to extract;
- a competitor answers the task more directly.
A citation gap can be a page-quality problem rather than a content-inventory problem.
Compare the pages that are winning
For each validated query, inspect the site Bing cites instead.
Compare:
- answer depth;
- freshness;
- named authorship;
- lists and tables;
- primary evidence;
- product detail;
- comparison clarity;
- original data;
- directness.
Do not copy the cited page.
Use the comparison to identify what makes the winning page easier to ground.
That is more useful than generic advice such as “write longer content.”
Do not treat Bing as a proxy for every AI system
Topics Portfolio is Bing Webmaster Tools data.
Do not treat Bing as a proxy for every AI system.
The report is not:
- Google AI Mode citation data;
- Google AI Overview data;
- ChatGPT citation data;
- Gemini citation data;
- Perplexity citation data.
Each system can use different indexes, retrieval methods and ranking processes.
A page can perform well in Bing’s AI experiences and poorly elsewhere.
The inverse can happen too.
That is why this report belongs inside a multi-surface AI visibility programme rather than becoming a universal visibility score.
NEMO’s Gemini 4 Argon monitoring framework follows the same principle: measure changes on the surface where they occur instead of assuming one AI system predicts another.
Use topic size to prioritise
Microsoft’s approximate topic ranges can still help distinguish:
- one isolated missing query;
- repeated gaps inside a topic;
- a broad cluster where competitors appear consistently.
A practical priority order can use four inputs:
Relevance × Business Value × Repeatability × Existing Page Potential
No complex formula is needed.
A commercially important query with repeatable competitor citations and an existing page that is nearly good enough should usually outrank a broad informational topic with little business value.
Small sites may not see the feature
Microsoft says sites with insufficient AI Performance data may not see Topics Portfolio.
That does not prove those sites never appear in AI answers.
It may simply mean Bing does not have enough observations to generate the report.
Likewise, seeing a large portfolio does not prove a site has more AI visibility than a smaller competitor.
The dataset reflects Bing’s available observations.
A simple seven-day workflow
The first week can stay small.
Day 1: Record the largest opportunity topics.
Day 2: Review 50 to 100 Opportunity Queries.
Day 3: Calculate Opportunity Query Validation Rate.
Day 4: Map validated queries to existing URLs.
Day 5: Calculate Fixable Citation Gap Rate.
Day 6: Compare your pages with cited beneficiaries.
Day 7: Select a small group of changes and save the baseline.
Then wait for enough fresh data before judging the effect.
Do not rewrite dozens of pages at once and then guess what changed the citation outcome.
The opportunity list is only the start
The important release is not another AI visibility score.
It is a first-party starting point for seeing where Bing finds other sites useful for grounding AI answers while yours is absent.
That can shorten the discovery stage.
But editorial judgment still matters.
Validate the query. Check intent. Confirm the beneficiary. Look for an existing page before commissioning a new one. Keep ranges as ranges.
The opportunity list is the beginning of the audit, not the conclusion.