Bing AI Visibility Insights: How to Use Intents, Topics, Citation Share, and Compare
Bing added Intents, Topics, Citation Share, and Compare to AI Performance reporting. Learn how SEOs should use the data without treating it like rankings.
Bing AI Visibility Insights Recap
Watch the 4-minute AI citation measurement walkthrough
A practical breakdown of Intents, Topics, Citation Share, Compare, and how to use Bing's AI Performance report without treating Citation Share like rankings.
TL;DR: Bing's June 16, 2026 update to AI Performance reporting gives SEOs a better way to measure AI citations. Intents, Topics, Citation Share, and Compare can turn AI visibility from screenshot hunting into a repeatable review workflow.
The warning matters as much as the feature list. Citation Share is a diagnostic metric. It is not an AI ranking, not traffic share, not a list of competitor domains, and not a quality score.
Bing's June 16, 2026 update to AI Performance reporting in Bing Webmaster Tools gives SEOs a better way to measure AI citations.
The update is useful because it moves the work away from random screenshots and toward a repeatable review process. You can now look at the context behind a citation, the subject cluster it belongs to, your share of the cited evidence for a grounding query, and how the pattern changes across time.
The four new preview capabilities are:
- Intents
- Topics
- Citation Share
- Compare
Used carefully, this can help content teams make better decisions. Used carelessly, it becomes another number in a dashboard that people overread.
The Reporting Gap AI Search Created
AI search created a measurement problem before most teams had a measurement process.
For a long time, the common method was simple: run a prompt, save a screenshot, count whether the brand appeared, and repeat the exercise later. That can help with spot checks, but it is weak evidence for planning content, reporting to clients, or deciding which pages to update.
Classic SEO has its own problems, but it gives us stable surfaces to inspect: impressions, clicks, crawl behavior, index status, query data, links, logs, and page performance. AI visibility has been much harder to review with the same discipline.
That is why the Bing update matters. It does not solve all AI search measurement. It gives publishers a first-party view into how their content appears as cited evidence across Bing-supported AI experiences.
If you use external prompt tracking too, keep using it with care. I covered the limits of prompt-only measurement in AI visibility prompt trackers. The Bing report is different because it starts from first-party citation data, not only from manually tested prompts.
What Bing Added on June 16, 2026
Bing launched the AI Performance report in Bing Webmaster Tools in February 2026. That first version showed how publisher content appeared as citations across Microsoft Copilot, AI-generated summaries in Bing, and select partner AI experiences.
The February report included citation totals, average cited pages, grounding queries, page-level citation activity, and trends across time.
The June 16 update adds context to that base layer.
| Layer | What it shows | Useful question | Caveat |
|---|---|---|---|
| Intents | The context category Bing assigns to a grounding query | Which kinds of AI experiences cite us? | This is not a classic rank-tracker intent tag |
| Topics | Related grounding queries grouped by subject | Which subject clusters create citation activity? | Preview labels can be general |
| Citation Share | Your citations divided by all citations for a query | How much of the cited evidence set do we receive? | This is not ranking, traffic share, or a quality score |
| Compare | A previous period overlaid on the current view | Did citation patterns move after an update? | Correlation does not prove cause |
That combination is the real story. Intents and Topics tell you where citations sit. Citation Share tells you how much of the cited evidence set you receive for a grounding query. Compare helps you review change across time.
For the Google side of this reporting problem, see Google search generative AI performance reports. Bing is giving publishers a more direct AI citation view than most teams currently have elsewhere, but it still needs to sit beside analytics, classic search data, and page-level review.
Grounding Is the Starting Point
Before the new fields make sense, you need to understand grounding.
In Bing's framing, grounding is the source material and web evidence an AI system uses to support and cite an answer. In plain SEO terms, it is the part of the answer process where the system relies on current web content instead of only the model's internal knowledge.
That does not make grounding a replacement for crawling, indexing, rankings, or traffic. It adds another surface to inspect.
A page can rank in classic search and have weak citation participation in AI answers. A page can receive AI citations and send little referral traffic. A page can be cited for low-value informational queries while missing the commercial or local contexts that matter to the business.
This is why the June update is useful. It helps separate those questions instead of treating every citation as equal.
Intents: The Query Context Behind a Citation
Intents classify grounding queries into categories.
Bing lists examples such as Informational, Commercial, Navigational, Learn and Solve, Research, Creation, and Local. The user-supplied Bing Webmaster Tools screenshot for this article also showed labels such as Live Event and Comparison in a weather-related example.
The field is useful because it helps answer this question:
Are we being cited in the AI query contexts that match business value, or only in low-value informational contexts?
Do not read Intents like a standard keyword intent tag. A grounding query belongs to an AI answer workflow. It may not map neatly to one keyword, one SERP, or one landing page.
Use the field to sort the report into practical buckets.
| Intent pattern | What to check | Possible action |
|---|---|---|
| Informational citations dominate | Are cited pages only answering basic definitions? | Add stronger next-step links where the topic supports the funnel |
| Commercial citations are weak | Do service, product, or comparison pages answer the query context clearly? | Improve decision criteria, proof, examples, and entity clarity |
| Local citations appear | Are local pages current, specific, and crawlable? | Tighten service area proof and local examples |
| Research citations appear | Are findings, data, tables, and caveats easy to extract? | Add clearer summaries and supporting evidence |
The win is not having an intent label. The win is noticing when your citation pattern does not match your business model.
Topics: Visibility by Subject, Not by One Query
Topics group related grounding queries into subject clusters.
Bing's example connects queries such as solar panels, solar energy efficiency, and residential solar installation into a topic like Solar Energy. That type of grouping is closer to how editorial strategy actually works.
Most sites do not improve AI visibility by chasing one prompt at a time. They improve hubs, pages, definitions, comparison assets, proof pages, internal links, and supporting content.
Topics can help you find where citation visibility clusters.
| Topic signal | What it may mean | What to inspect |
|---|---|---|
| High citation activity, weak business value | Visibility exists, but not where it matters most | Internal paths from cited pages to stronger service or conversion pages |
| Low citation activity, high business value | The site may lack clear evidence or coverage | Content gaps, stale pages, thin proof, poor structure |
| Very general topic label | The preview classification may be too loose | Grounding queries and cited pages before planning edits |
| One page carries the topic | One asset may be doing too much work | Supporting pages, examples, FAQs, and internal links |
Bing notes that topic labels can be general during preview, especially for specialized sites. Treat the labels as directional. They help you decide where to investigate, not what to rewrite automatically.
This connects with the wider AI search documentation issue. Google's own guidance still points teams back to useful, crawlable, well-structured pages. I covered that in Google AI search optimization documentation. Bing's topic data gives you another way to find where that page-level work may be needed.
Citation Share, Explained Without Hype
Citation Share is the metric most likely to be misread.
Bing calculates it like this:
Citation Share = citations attributed to your site / all citations shown across all sites for the same grounding query
In plain English, it is your share of the cited evidence set for a specific grounding query.
That is useful. It tells you more than whether your site appeared. It helps you see where your visibility is concentrated, where it is fragmented, and which query groups deserve page-level review.
It does not tell you:
- Your AI ranking position
- Your traffic share
- Which competitor domains received the remaining citations
- Whether Bing assigned your page a quality score
- Whether one content update caused the change
The safer wording is "share of cited evidence for this grounding query."
| Safe use | Unsafe use |
|---|---|
| Compare Citation Share across important grounding queries | Call it an AI ranking |
| Segment it by intent and topic | Report it as traffic share |
| Pair it with analytics and referral data | Estimate revenue from Citation Share alone |
| Use it to prioritize page reviews | Treat low share as a content quality grade |
| Watch it before and after updates | Claim a single edit caused the movement |
This matters for client reporting. A clean percentage can feel more precise than it really is. The field is valuable because it creates a relative citation signal, but the limits must travel with the number.
For more context on why citation presence and business impact are not the same thing, see ChatGPT referral traffic tracking and AI citation ranking factors.
Compare: The Feature That Turns a Report Into a Test
Compare lets publishers overlay a prior period onto the current reporting view. You can use current 30 days versus prior 30 days, or another date range.
This is where the report becomes operational.
Without Compare, you can say a page was cited. With Compare, you can ask whether citation patterns changed after a content refresh, during seasonality, after a shift in user demand, or after a wider model or partner change.
The discipline is causal humility. Compare can show movement. It cannot prove why the movement happened.
AI citation patterns can shift because of user behavior, model changes, freshness signals, partner refresh cycles, and changes across the web. A rising Citation Share after an update is a lead to investigate. It is not proof by itself.
A Practical AI Visibility Audit Workflow
Here is how I would use the June 16 Bing update for a publisher, ecommerce site, SaaS company, or service business.
1. Export the Data
Start with the export, not the screenshot.
You want grounding queries, intents, topics, cited pages, citations, Citation Share, and date ranges. If Bing gives you query and page views, keep both.
2. Filter for Business-Relevant Intents
Do not chase the largest citation count first.
Start with intent. A publisher may care about research and learning queries. A local service business may care about local and comparison contexts. An ecommerce site may care about shopping, comparison, and decision-support contexts.
3. Group by Topic
Group rows by topic to see whether visibility lines up with your real content architecture.
A topic with many citations and weak page alignment may mean you have demand but no clean hub. A topic with few citations and high commercial value may deserve a content gap review.
4. Map Queries to Pages
For each important query cluster, inspect the cited page.
Ask:
- Does the page answer the intent directly?
- Is the main answer easy to extract?
- Are definitions, steps, comparisons, and caveats clear?
- Is the page current?
- Does it include real proof, examples, or data?
- Does it link to the next useful page?
5. Improve Only Pages That Match the Query Context
Do not update a page just because Citation Share is low.
Update when the page is a legitimate match for the intent and topic, and when the content has a fixable weakness.
Good edits include:
- Clearer answer blocks near the top of relevant sections
- Better comparison tables
- Dated freshness notes where freshness matters
- Examples from real workflows
- Stronger author or business proof
- Internal links to related guides, tools, or service pages
- Concise FAQ answers for recurring grounding-query patterns
Weak edits include:
- Stuffing "AI visibility" into every page
- Adding unsupported claims
- Rewriting a page for a query it should not answer
- Treating one topic label as a full content strategy
6. Use Compare After the Next Measurement Window
After the update has had time to be crawled and appear in AI reporting, use Compare.
Look for movement by query, intent, topic, cited page, and Citation Share. Then validate with other data:
- Bing search performance
- Analytics landing page data
- Referral reporting from AI surfaces where available
- Google Search Console data
- Crawl and index checks
- Lead, trial, or revenue reporting where relevant
This gives you a useful review loop:
| Step | Output |
|---|---|
| Export AI Performance data | Working sheet with queries, pages, intents, topics, citations, and Citation Share |
| Mark priority intents | Filtered list of contexts that matter to the business |
| Cluster by topic | Shortlist of subject groups to inspect |
| Review low-share, high-value rows | Page-level improvement notes |
| Pick 3 to 5 page updates | Small test set with dates recorded |
| Use Compare later | Before-and-after review with caveats |
| Pair with other data | More complete visibility, traffic, and outcome view |
The goal is not to "optimize for Bing AI" in a vague way. The goal is to choose a topic, inspect the evidence, improve pages that deserve the query context, and measure what changes.
For teams that need help turning this kind of report into a repeatable process, see AI SEO and technical SEO advisory.
What This Still Does Not Tell You
The June update is useful, but it leaves gaps.
It does not show every AI surface on the web. It does not replace Google Search Console. It does not show competitor domains behind Citation Share. It does not turn citations into traffic. It does not solve attribution from AI answers to leads or revenue.
It also does not tell you whether a user trusted the answer, clicked a source, copied advice, visited later, or converted through another channel.
That matters because citation visibility and business impact are related, but they are not the same thing. Use Bing Webmaster Tools data with analytics, referral reporting, Bing search performance, crawl checks, and classic search data.
What SEOs Should Do This Week
If you have access to the updated Bing Webmaster Tools preview, run one focused pass:
- Open the AI Performance report and confirm the new fields are present.
- Export grounding queries and pages.
- Filter the export by intents that matter to the business.
- Group the rows by topic.
- Sort priority groups by Citation Share.
- Pick a small set of pages where the query context and page purpose match.
- Improve evidence, structure, examples, freshness, and internal links.
- Record the update date.
- Use Compare after the next reporting window.
- Validate movement with analytics, search performance, and referral data.
Keep the first pass small. You are building a measurement habit, not a sitewide rewrite plan.
FAQ
Is Citation Share a new AI ranking factor?
No. Bing describes Citation Share as observational. It shows the share of citations attributed to your site for a specific grounding query. It is not a ranking position or a quality score.
Does Citation Share show AI traffic share?
No. Citation Share is based on citations in the reported AI answer context. It does not tell you traffic, clicks, leads, or revenue.
Can I see which competitors received the rest of the citations?
No. The report does not expose competitor domains. Treat Citation Share as a relative citation signal, not a named competitor export.
What should I do if a topic has low Citation Share?
Start with diagnosis. Check whether the cited or target page matches the query intent, whether it is current, whether the answer is easy to extract, and whether the page has enough proof. Low share can point to a content gap, weak structure, stale information, or a query where AI systems use many sources.
Does Compare prove that my content update improved AI visibility?
No. Compare helps you observe change across time. It does not prove cause by itself. Use it with crawl data, Bing search performance, analytics, referral reporting, and a clear update log.
Should this replace AI visibility prompt tracking?
No. It can reduce dependence on manual screenshot checks for Bing-supported AI experiences. Prompt tracking can still help with spot checks across other answer engines, but Bing's report gives a better first-party measurement layer for its own supported surfaces.
Does this replace Google Search Console?
No. It covers Bing-supported AI citation behavior. Keep using Google Search Console, analytics, crawl data, and referral tracking beside it.
Final Take
Bing's update is useful because it gives SEO teams a cleaner way to ask better questions.
Intents tell you the query context. Topics show the subject cluster. Citation Share shows your portion of the cited evidence set for a grounding query. Compare lets you inspect change across time.
The right question is not "what is our AI ranking?"
The right question is: "For the intents and topics that matter, are our pages being used as evidence, and can we improve that pattern without pretending the data says more than it does?"
