Microsoft Clarity has added branded and non-branded grounding-query segmentation to the AI Citations dashboard. The feature went live on August 3 and received a detailed new analysis from Search Engine Journal on August 10. Marketing teams can now separate retrieval queries that mention the brand from broader category or problem queries. Share of Authority is also broken out for the two groups.
This closes an important blind spot in reports that reduce AI visibility to one citation total. Retrieval for a named brand is not the same as being selected as a source for a broader topic. The first can show that the brand entered the research process. The second may show that its content earned category discovery. Neither one proves user intent, a click, or revenue by itself.
What exactly changed?
The query cards in Clarity's Citations dashboard now label relevant grounding queries as branded. Users can filter branded and non-branded queries and compare visibility. Share of Authority is also segmented by query type. Microsoft says this makes it easier to distinguish direct brand visibility from category discovery.
A grounding query is not necessarily the original prompt typed by the user. An AI system can generate its own searches to retrieve supporting information. A user may ask, “Which tools help explain visitor behavior?” The system can then search for specific products by name. Clarity may classify those retrieval queries as branded even though the user's prompt contained no brand.
This text-free Fark Studio diagram compares retrieval focused on one distinctive brand object with retrieval that selects sources across a wider category. It is not a Clarity interface.
Share of Authority represents the share of page citations attributed to a domain compared with other sources in included queries where that domain received a citation. It should not be read as a brand's total share of the AI search market. Queries where the domain received no citation or falls outside coverage may not enter the calculation in the same way. The metric is a trend and competitive signal, not a market census.
Why citations, visits, and conversions must stay separate
A page citation does not mean a user clicked the source. The answer may be sufficient, the link may be difficult to notice, or the page may have been used only during grounding. Clarity tracks AI referral sessions separately. Business impact still needs analytics, CRM, and sales data.
Teams therefore need three stages. First, visibility: which page received a citation for which grounding query? Second, traffic: did that AI surface send a website session? Third, value: did the visit become a lead, sale, pipeline opportunity, or another business result?
This Fark Studio flow separates citation visibility, a website visit, and a business outcome. It shows that value can drop between stages and does not claim automatic revenue growth.
A brand can gain citations while referral traffic remains flat. That may indicate influence inside answers, but it does not prove commercial impact. Referral traffic can also rise while citation share falls, because a few high-volume visits do not represent the full visibility landscape.
Who is affected?
SEO and content teams can use non-branded queries as a stronger signal of category authority. Product documentation may perform well for branded retrieval, while guides, research, and comparison content can earn citations for non-branded topics. Separating the groups creates a more useful investment decision.
Brand and communications teams can see whether third-party sources dominate branded citations. If an AI system looks for information about the company but relies on review platforms, publishers, or forums instead of the official site, pricing, product, policy, and reputation may be shaped externally.
Performance and CRO teams should not confuse AI referral traffic with citation visibility. Strong citations and weak visits can justify a review of page titles, link context, and the next step offered to users. Strong visits and weak conversion can indicate a landing-page or proposition problem.
Executives and agencies face a reporting risk. Turning the feature into one “AI SEO score” erases the different meanings of branded and non-branded retrieval. Clarity provides a representative view across supported Microsoft and partner AI experiences, not a complete record of every surface from ChatGPT, Gemini, Claude, Perplexity, and others.
Coverage and methodology limits
Microsoft's methodology says the dashboard aggregates grounding and citation activity across supported AI experiences. Data is refreshed daily with a short processing delay. Very low volume can be excluded. Page and grounding-query totals can differ, and metrics may be refined as more information is processed.
Microsoft grounding supports its own AI experiences and partner platforms that use Bing for retrieval. Microsoft does not publish a complete platform list. No citation in the report does not prove that no AI system on the web used the page. This is a sampled trend tool.
Branded classification has edge cases. A brand may also be a common word. Companies use abbreviations, product names that differ from the corporate name, and queries mentioning several brands. Monthly reporting needs manual review of representative examples.
What should brands do now?
1. Verify the domain. A project administrator can use the Clarity tracking code, Google Search Console, or Bing Webmaster Tools. Citation reporting should not enter the operating process before verification is complete.
2. Build a branded dictionary. List the brand name, misspellings, product names, former names, and common abbreviations. Sample Clarity's classification against that list.
3. Create non-branded topic clusters. Organize real customer problems into categories, use cases, and decision questions. Do not simply copy a high-volume SEO keyword list.
4. Save a monthly baseline. Record branded citations, branded Share of Authority, non-branded citations, non-branded Share of Authority, AI referral sessions, and AI-sourced conversions for the same period.
5. Review the page level. Identify which URLs drive the movement. If old, inaccurate, or commercially weak pages earn citations, visibility growth can also create an information risk.
6. Classify competing sources. Review which domains receive citations in the same grounding queries. A competitor product page, independent publication, and user forum should not sit in one undifferentiated competitor bucket.
7. Report traffic separately. Do not present a citation-share increase as a website-traffic or sales increase. Each stage needs its own metric.
8. Add conversion quality. Compare lead quality, sales acceptance, revenue, and acquisition cost from AI referral sessions with other channels.
9. Choose a content action. Weak branded visibility may call for clearer product, pricing, policy, support, and company pages. Weak non-branded visibility may call for original research, category guidance, comparisons, and third-party authority.
10. Put a scope note on every report. Describe the data as a Microsoft Clarity sample across supported AI experiences, not the entire AI search market.
This Fark Studio illustration shows an analyst building a monthly scorecard from separate brand, category, citation, visit, and outcome signals.
Where should teams wait?
Do not call one week a trend. Processing delay, low volume, and classification error can create large percentage changes on small sites. Review several periods and sample the underlying queries before changing strategy.
High branded Share of Authority does not prove category demand has been won. The brand may already be present in the retrieval query. A non-branded increase does not automatically mean purchase intent because a grounding query may differ from the user's original prompt.
Do not infer zero visibility on an unsupported AI platform. Measurement tools have different coverage and methods. Preserve platform-level limitations instead of combining them into one supposedly definitive number.
Fark Studio perspective
The new split makes AI visibility reporting more honest. Being researched by name and being selected for expertise without a brand mention are different growth paths. A good report shows both and keeps citations separate from traffic and revenue.
At Fark Studio, SEO, performance marketing, and corporate web should share one AI visibility scorecard. Content verifies the cited source, analytics verifies the visit, and sales verifies the outcome.
If you want to measure branded and non-branded AI visibility, strengthen the right source pages, and connect the result to CRM data, Fark Studio can run an AI visibility audit.
Sources
Microsoft Clarity Blog, See Branded and Non-Branded AI Queries in the Clarity Citations Dashboard, August 3, 2026. Primary feature announcement.
Microsoft Clarity Blog, Understanding Your Influence in AI Answers with Microsoft Clarity, February 17, 2026, current methodology. Primary source for coverage and limitations.
Search Engine Journal, Microsoft Clarity Now Separates Branded And Non-Branded AI Citations, August 10, 2026. Secondary marketing analysis and implementation limits.
