Microsoft Clarity Shows Brand Visibility Is Moving Below the Click
Microsoft Clarity's AI citation split shows why brand visibility now lives below the click.
Microsoft Clarity's latest AI visibility updates show that brand measurement is moving below the click. The useful question is no longer only whether AI systems cite a page. It is whether those citations accompany branded grounding queries or broader, non-branded grounding queries—the searches AI systems run to build an answer, not necessarily the prompts people typed.
Microsoft Clarity turns AI citations into a brand discovery split
Microsoft Clarity now separates branded and non-branded grounding queries associated with AI citations, giving marketers a clearer read on brand-specific retrieval and broader category discovery. Microsoft described the shift in an August 12, 2026 Microsoft Advertising post about how humans and AI "find and choose your brand" with Clarity, after earlier Clarity releases added AI citation reporting and branded query segmentation. The core move is simple: Clarity measures AI citations separately from AI referral traffic and now segments citation analysis by branded and non-branded grounding queries.
That distinction matters because a brand can look healthy in aggregate while still missing the category conversation. Microsoft Clarity's own blog says branded and non-branded AI queries are not the same kind of visibility: branded grounding queries name the brand directly, while non-branded grounding queries cover broader topics used to build an answer.
Search Engine Journal's coverage of the Clarity release framed the change as branded and non-branded query labels, filters, and Share of Authority breakdowns inside the AI Citations dashboard. The operational point is bigger than a dashboard feature. A CMO can compare citations associated with brand-specific retrieval against citations associated with broader topics, then inspect which pages and competing domains appear in each group.
AI visibility below the click changes the CMO dashboard
The click is becoming an incomplete unit of brand visibility because AI systems can retrieve, cite, and summarize a page without producing a visit. Microsoft Clarity's May 2026 announcement said AI-generated answers are becoming one of the ways people discover information online, and its AI Citations feature is designed to show when a site is cited in AI-generated responses even when standard analytics undercount that exposure.
That changes how marketers should read performance. A page can create value before a user lands on the site. It can inform an answer, support a comparison, or make a brand look credible in a generated recommendation. Microsoft's AI Citations methodology describes a representative view of grounding and citation activity across supported Microsoft and partner experiences that rely on Bing retrieval—not a complete log of every AI reference. Very low volumes may be excluded, and totals can differ between page and grounding-query views. These are source-level signals, not just session-level signals.
The useful dashboard now has two layers:
| Measurement layer | What it answers | Why it matters |
|---|---|---|
| Branded AI citations | Do AI systems cite our pages for grounding queries that name our brand? | Shows brand-specific retrieval and citation visibility. |
| Non-branded AI citations | Do AI systems cite our pages for broader grounding queries that do not name our brand? | Shows broader-topic discovery and competitive source visibility. |
| AI referral traffic | What percentage of site sessions comes from AI assistants? | Shows AI-referred visits, not a citation-to-visit conversion rate or a measure of zero-click influence. |
| Share of authority | What percentage of citations goes to our domain versus other domains within the same queries where our domain appeared? | Shows relative citation share within that included query set, not the entire AI search market. |
This is why the Microsoft Clarity signal belongs in the broader AI visibility conversation. Visibility is no longer a ranking position on a search results page. It is a brand's ability to be retrieved, cited, and trusted inside machine-generated answers.
Branded AI citations are not the same as category authority
A branded citation records brand-specific grounding; a non-branded citation records citation visibility during broader retrieval. That distinction is the strategic value of Microsoft's branded versus non-branded split. An AI system can generate branded grounding queries from a prompt that never names the brand. A non-branded grounding query identifies a broader retrieval topic, not necessarily the wording of the user's prompt. It is the brand's source architecture.
That is the measurement gap most teams miss. They see AI traffic and ask whether it is up or down. A better question is whether the brand's pages are being cited for broader, non-branded grounding queries as well as branded ones. That split exposes different retrieval contexts, not what the user already knew.
The same logic applies to citation architecture. Pages need direct answers, source-backed claims, clear entity language, and crawlable evidence because machines are selecting passages and sources before users decide what to click. A homepage built for human persuasion is not enough if the machine cannot extract the company, category, proof, and relevance from the page.
What brands should change after the Microsoft Clarity update
The practical response is to map AI citations by query intent, then rebuild the weak source layer. A brand should not treat every AI citation as equal. Branded, non-branded, comparison, pricing, alternatives, support, and category-definition queries all tell different stories.
For a CMO, the first pass should be mechanical:
- Separate branded from non-branded AI citation queries.
- Identify which pages AI systems cite for each query class.
- Compare cited pages against the claims the brand wants machines to repeat.
- Repair pages that are cited but vague, outdated, or weakly sourced.
- Build new source pages only where no existing page can answer the category query cleanly.
The most important move is the fourth one. If an AI system is already retrieving a page, the brand has a live source surface. The work is not to publish more noise. It is to make that page more extractable: stronger opening answer, clearer entity attribution, better proof, cleaner comparison structure, and internal links that help machines resolve the brand's role.
This is where Machine Relations becomes a useful frame. Jaxon Parrott has described Machine Relations as the shift from optimizing only for human-mediated discovery to making brands legible, retrievable, and credible to machine-mediated discovery systems. In the Microsoft Clarity context, that means the measurement layer has to connect back to the source layer. If the dashboard shows weak non-branded citations, the problem is usually not the dashboard. It is the brand's source architecture.
FAQ
What did Microsoft Clarity add for AI brand visibility?
Microsoft Clarity has added AI citation reporting and branded versus non-branded grounding-query segmentation, helping site owners see whether their pages are cited alongside brand-specific searches or broader topic searches generated by AI systems. Microsoft explains the AI citation feature in its Clarity documentation and blog updates.
Why do branded and non-branded AI citations matter?
Branded AI citations show that a site's pages were cited alongside grounding queries that name the brand. Non-branded AI citations show citation visibility alongside broader grounding queries. It is the brand's source architecture.
Is Microsoft Clarity enough to measure AI visibility?
Microsoft Clarity is useful for source-level AI citation signals, but it should not be the only measurement layer. Brands still need prompt testing, source-quality audits, competitive citation tracking, and post-click analytics to understand whether AI visibility is changing demand, perception, and pipeline.
How should a brand improve weak non-branded AI visibility?
Start with the pages AI systems already retrieve, then strengthen their answer-first structure, entity clarity, source citations, comparison tables, and internal links. If there is no page that answers the category query directly, build one.