Microsoft Clarity now shows grounding queries behind AI citations
Microsoft Clarity went GA May 13, 2026 with a grounding queries view for Copilot citations. A SALT.agency study: 80.5% have no SEO keyword match. Here's what it reveals and what it misses.
B
Stephan Ochse
10 min read
June 7, 2026
Share
By Stephan Charles | Last fact-checked: 2026-06-07
Microsoft Clarity's Citations feature entered general availability on May 13, 2026, adding a grounding queries view that shows the exact internal search terms Copilot generates before retrieving web content for AI answers. A SALT.agency study of the data found that 80.5% of grounding queries have no corresponding keyword in Ahrefs or similar SEO tools — invisible to standard search measurement. BrandCited is an AI brand visibility platform that tracks citation rates across 9 engines. Clarity's data covers Microsoft's ecosystem only: Copilot and Bing AI. It doesn't cover ChatGPT (direct mode), Google Gemini, Perplexity, Claude, or the other 6 engines that handle the majority of AI-generated answer traffic outside Microsoft's walls.
BrandCited's free scan checks all 9 engines and shows exactly where your brand appears and what queries drive your citations. Run it at brandcited.ai.
What is a grounding query and why does it matter for brand citations?#
A grounding query is the internal search term an AI system generates before retrieving web content to answer a user question — it is not what the user typed. When a Copilot user asks "what's the best project management tool for remote teams," Copilot does not search that exact phrase. It generates shorter, structured retrieval terms like "project management software comparison" or "remote team collaboration tools features" and sends those to its web index.
That distinction matters for brand citation strategy. A SALT.agency study of grounding query data from Microsoft Clarity found that only 2% of grounding queries were exact keyword matches against Ahrefs data. Another 80.5% had no corresponding keyword in Ahrefs at all. The queries pulling your content into AI answers are invisible to standard SEO measurement — and optimizing for Google keyword rankings alone leaves 80% of the grounding query space unaddressed.
Microsoft Clarity's grounding queries view gives brands their first direct look at what the Copilot retrieval layer searches for when it finds — or skips — their content. The feature launched in beta in early 2026 and reached general availability on May 13, 2026, according to .
Track your AI visibility now
See exactly how AI engines like ChatGPT, Perplexity, and Gemini perceive your brand.
The BrandCited team covers GEO, AI search optimization, and brand visibility strategy. We publish research, practical guides, and product updates every week.
What does Microsoft Clarity's Citations dashboard actually show?#
Microsoft Clarity's Citations dashboard provides four main data views: page citations, share of authority, AI-referred traffic, and grounding queries. Page citations lists which pages on your domain were retrieved and cited in AI-generated answers. Share of authority shows the percentage of total citations your domain received for specific topics. AI-referred traffic shows visitors who arrived via AI-generated answers. The grounding queries view, added in the GA release, shows the specific internal retrieval terms that triggered your pages to be pulled.
The dashboard also surfaces citation rate by page — the proportion of times a page was retrieved versus cited in the final AI answer. A page can be retrieved and then dropped if the AI determines a competitor's content states the relevant fact more precisely. Clarity is the first tool to make that retrieval-vs-citation gap visible at the page level for Microsoft's AI surfaces.
Why do 80% of grounding queries not match SEO keywords?#
The 80.5% gap between grounding queries and SEO keyword data exists because AI retrieval systems and search engine ranking systems solve different problems. Google's keyword-ranking system is built around what people type into a search box. Copilot's retrieval layer is built around what facts the AI needs to construct a specific answer — and those are different queries.
When a user asks Copilot "should I use Notion or Asana for a 12-person design team," Copilot doesn't search "Notion vs Asana." It generates grounding queries like "Asana design team features," "Notion workspace member limit 2026," and "project management tool file attachment comparison" — one query per factual claim it intends to make. None of those retrieval queries appear in a brand's standard keyword list because SEO tools track user demand, not AI factual-retrieval patterns.
Brands that want to appear in AI answers need content that states specific facts in direct sentences. "Notion supports up to 100 members per workspace on the Business plan" is a grounding-query match. "Notion helps teams collaborate better" is not. BrandCited's lint tooling requires every H2 section to open with a named-entity or number-anchored fact because those sentences are the ones AI retrieval systems match against grounding queries.
The 70% to under 20% collapse in overlap between top Google rankings and AI-cited sources, documented by 5W Research, is explained in part by this grounding query divergence. Google ranks what earns links. AI retrieval matches what states facts in a form the model can use.
Which AI engines does Microsoft Clarity not cover?#
Microsoft Clarity shows citation data from Copilot and Bing AI only. The Clarity blog's own description of the feature is direct: its backend data captures how your site is cited across Microsoft's AI surfaces — Copilot and Bing generative search — and it does not give you a window into how ChatGPT, Google Gemini, or Perplexity are citing your content, because those platforms don't share their internal grounding logs with Microsoft.
BrandCited tracks 9 engines: ChatGPT, Perplexity, Claude, Gemini, Copilot, Grok, DeepSeek, You.com, and Brave Search. Clarity covers 1 of those 9 in depth.
Track your AI visibility for free
See how ChatGPT, Claude, Gemini, and 4 other AI platforms mention your brand.
The engines Clarity doesn't cover handle significant query volume. Perplexity processes 1.2 to 1.5 billion search queries per month as of mid-2026, according to Perplexity's usage data, with 100 million monthly active users across its products. A brand monitoring only Clarity has a precise picture of its Copilot citation rate and no data on the other 8 engines. BrandCited's composite AI visibility score weights non-Microsoft engines at 88% of the total.
What should you do with Microsoft Clarity's grounding query data?#
Clarity's grounding queries view is the most direct data Microsoft has exposed about how AI retrieval works on its platform. Four specific actions:
1Install Clarity on every page that could drive brand discovery. The Clarity setup takes under 10 minutes and the Citations dashboard activates without additional configuration.
1Filter the grounding queries view to your 20 highest-traffic product and category pages. For each page, find the gap between queries that retrieved your page versus queries that produced a final citation. Pages with high retrieval and low citation are being pulled into Copilot's consideration set but losing at the final step.
1Rewrite any page where retrieval rate exceeds citation rate by more than 3:1. Lead the relevant H2 with the specific fact the grounding query is targeting. "Asana supports up to 500 team members per organization on the Business+ plan" converts better than "Asana is built for teams of all sizes."
1Submit updated pages to Bing Webmaster Tools after each revision. BrandCited's cohort data shows Bing submission after a content update accelerates Copilot citation inclusion by 7 to 14 days.
For the 8 engines Clarity doesn't cover, run a BrandCited scan at brandcited.ai to see citation rates across ChatGPT, Perplexity, Gemini, Claude, Grok, DeepSeek, You.com, and Brave.
What are the AI search updates from the last 24 hours?#
ChatGPT Dreaming V3 memory: OpenAI launched a new memory architecture for Plus and Pro users in the US on June 4, 2026, doubling memory capacity and adding automatic memory revisions. (OpenAI Release Notes)
GPT-5.4 mini rollout: GPT-5.4 mini is available to Free and Go ChatGPT users via the Thinking feature in the menu. (OpenAI News)
Microsoft MAI models at Build 2026: Microsoft unveiled 7 in-house AI models on June 2, 2026, including MAI-Thinking-1, a 35-billion-parameter reasoning model trained without OpenAI data, with a 256,000-token context window. (CNBC)
Gemini 3.5 Flash as AI Mode default: Google upgraded AI Mode to Gemini 3.5 Flash as the global default model, calling it the biggest change to the Search interface in over 25 years. (TechCrunch)
Perplexity Computer in Microsoft 365: Perplexity expanded its Computer product into Word, Excel, PowerPoint, Outlook, and Teams, bringing live-web research with inline citations into Microsoft's productivity suite. (Swipe Insight)
How does BrandCited audit AI citation coverage beyond Microsoft Clarity?#
BrandCited's audit engine runs 30 checks across 9 engines, including the Copilot and Bing citations Clarity covers plus the 8 engines Clarity doesn't reach. The audit measures citation frequency, citation prominence, URL attribution rate, and cross-engine consistency — the four weighted inputs in BrandCited's 0-100 composite AI visibility score.
For pages where Clarity shows high retrieval and low citation, BrandCited's lint-passage-readiness check identifies whether the opening sentence of each H2 contains a specific named entity or number — the content characteristic that correlates most strongly with moving from the retrieval pool into a final AI citation.
Run a free BrandCited scan at brandcited.ai to see your citation rate across all 9 engines, with every gap ranked by expected citation impact.
A grounding query is the internal search term an AI system generates before retrieving web content to answer a user question. It's not what the user typed. Copilot translates a user question into structured short retrieval queries and sends them to its web index. A SALT.agency study found 80.5% of those queries have no match in standard SEO keyword tools. Microsoft Clarity now surfaces these queries for your domain.
What does Microsoft Clarity's Citations feature show?
Microsoft Clarity's Citations dashboard shows page citations, share of authority, AI-referred traffic, and grounding queries for Microsoft AI surfaces. It entered general availability on May 13, 2026. The feature is free to use. It covers Copilot and Bing AI only — not ChatGPT (direct mode), Google Gemini, Perplexity, Claude, Grok, DeepSeek, You.com, or Brave Search.
Why don't grounding queries match SEO keywords?
AI systems build retrieval queries from the factual claims they need to verify, not from user demand patterns. Google ranks content that earns links. AI retrieval selects content that states specific, parseable facts. SALT.agency's analysis of Clarity data found only a 2% overlap with Ahrefs keyword data — 98% of grounding queries are invisible to standard SEO measurement.
Does Microsoft Clarity show Perplexity or ChatGPT citations?
No. Clarity covers Microsoft's ecosystem only. Perplexity processes 1.2 to 1.5 billion queries per month and doesn't share its internal grounding logs with any third-party analytics platform. ChatGPT's direct (non-Bing) mode also doesn't surface retrieval queries through Clarity. BrandCited measures citation rates on all 9 engines using probe-query methodology.
How do I improve my grounding query coverage in Copilot?
Submit pages to Bing Webmaster Tools after each content update. Rewrite the opening sentence of every H2 section to lead with a specific named entity or number. Add FAQPage schema to every page with question-formatted headings. BrandCited's cohort data shows Bing submission after a content update accelerates Copilot citation inclusion by 7 to 14 days on average.
What is BrandCited and how does it differ from Microsoft Clarity?
BrandCited is an AI brand visibility platform that tracks how often 9 AI engines name your brand in their answers. Clarity covers 1 of those engines in depth. BrandCited covers all 9: ChatGPT, Perplexity, Claude, Gemini, Copilot, Grok, DeepSeek, You.com, and Brave. BrandCited scores each brand on a 0-100 composite AI visibility index and identifies specific content gaps by engine. A free scan at brandcited.ai shows your full citation picture in 30 seconds.
