Microsoft Clarity for SEO

Microsoft Clarity is a free behaviour analytics tool: session recordings, heatmaps, and engagement metrics, installed with a JavaScript tag. Those features are conversion work rather than SEO work. What makes Clarity worth an SEO practitioner’s attention is the AI Visibility section Microsoft added through 2026, which reports citation activity in AI-generated answers, topic-level opportunity analysis, and server-side crawler data. All of it is free, which changes the entry price for measurement that dedicated platforms charge for.

The three dashboards do not share a methodology, and the difference matters more than the feature list. Citation reports activity Microsoft observed. Topic Insights runs prompts you define through a model and reports what comes back. Bot Activity reads your own server logs. Reading all three as one number is the fastest way to draw a wrong conclusion.

What does the Clarity Citation dashboard measure?

The Citation dashboard reached general availability on 13 May 2026.¹ It reports five things:²

• Page citations. The number of times pages from your domain were referenced in AI-generated answers during the period.
• Share of authority. Your citations as a percentage of total citations from all domains.
• AI referral traffic. The percentage of sessions on your site that arrived from AI assistants.
• Grounding queries. The queries AI systems used to retrieve your content before generating an answer. Microsoft is explicit that these “may differ from the exact phrases users type”.²
• My cited pages. Which URLs were referenced, with citation counts and the queries attached to each.

Grounding queries are the most useful of the five, and the least like anything in Search Console. A grounding query is the retrieval string the model generated, not the human question that prompted it. That makes it a direct view of how a system decomposed a request into sub-queries, which is the mechanic covered in query fan-out. Seeing the actual retrieval strings your content answers is a better guide to passage structure than guessing from keyword tools.

Two caveats belong next to the numbers. Share of authority is calculated daily and only counts days on which your domain was cited for a given query, which Microsoft acknowledges “may lead to higher Share of authority values compared to broader calculations”.² Read it as a relative trend, never as an absolute market share. More broadly, the dataset is described as “a representative view of grounding and citation activity rather than a complete log”, with low volumes excluded and totals varying between views, “designed for trend analysis and comparative insights over time, not precise accounting”.³

How is Clarity different from AI visibility tools?

Platforms like Profound, Semrush’s AI Visibility Toolkit, and Peec AI work by submitting a prompt set through AI platform APIs and parsing the responses for brand mentions. That is a simulation: you choose the prompts, and the results tell you what a model returns for prompts of your choosing.

The Clarity Citation dashboard inverts this. It reports grounding and citation activity Microsoft observed for your verified domain, so the query set is not yours to choose. You find out which queries actually pulled your content in, including queries you would never have thought to track.

Neither view is complete, and they fail in opposite directions. A prompt panel can measure a competitive topic where you have no presence at all, because it does not depend on you being cited to run the test; Clarity cannot show you a topic you are absent from, because absence produces no observation. Clarity, meanwhile, surfaces demand you did not anticipate, which no prompt list will ever contain. Running one and calling it AI visibility measurement leaves a specific, predictable blind spot.

Microsoft markets the distinction hard, contrasting its “real query-level data” and “actual AI generated citations” against competitors it characterises as offering “third-party scraping & estimates” and results “approximated from simulations”.⁴ The observation is fair as far as the Citation dashboard goes. It sits awkwardly with the rest of the product, because Clarity now does both: Topic Insights is a prompt panel bolted onto an observation tool, running the same kind of simulation the comparison table dismisses. Know which dashboard you are reading before you act on it.

What are Topic Insights and Query Topics?

Topic Insights launched on 9 July 2026.⁵ You define topics, prompts, and competitors, and it reports four things per topic: your visibility (citation frequency and authority share), your influence (how much your content contributed to answers), your competition (which domains appear alongside you and where they outperform), and your opportunities (prioritised content gaps).⁵

Read the disclaimer before you read the output. Microsoft states it is “using GPT-5.3 with an allowance of 10 reports per week per project and are leveraging WebIQ as our search grounding layer”.⁵ So Topic Insights is a simulation, and specifically a simulation grounded by Microsoft’s own retrieval layer rather than by the platform any given reader cares about. It answers “what would a GPT-5.3 answer grounded on Web IQ cite for this prompt”, which is a reasonable proxy and not the same question as “what does ChatGPT cite”. The ten-reports-per-week cap also rules it out as a monitoring tool; it is for periodic analysis.

Query Topics is the observational counterpart, in beta. It automatically groups the grounding queries you already earn citations for into a smaller set of generated topics.⁶ Where Topic Insights asks where the opportunities are, Query Topics asks where you are already earning visibility. Query Topics is the one to trust for reporting, because it is built from observed activity.

What does Bot Activity show?

Bot Activity is the dashboard least discussed and, for technical SEO, arguably the most useful. It reports which automated systems request your site, how often, and which paths they hit, broken down by bot operator, AI request share, activity category (search, data services, developer tools, AI assistants), and path requests.⁷

The data source is what makes it interesting: “real server-side logs collected through supported CDN and server integrations”, not the JavaScript tag.⁷ That is log file analysis with the parsing and storage handled for you, and it is the cheapest route to seeing AI crawlers hitting your site. Enabling it requires connecting a supported CDN or edge provider: Cloudflare, Fastly, Amazon CloudFront, Azure Front Door, and Akamai are all documented.⁸ Bot Activity is off until that connection exists, and WordPress sites running the current Clarity plugin get it by default.⁹

Two things to know before enabling it. Log access, data transfer, and processing charges are billed by your CDN provider, not by Microsoft, so the dashboard is free but the pipe feeding it may not be.⁷ And Microsoft states plainly that bot activity “doesn’t indicate that content was retrieved, grounded, cited, or surfaced in AI generated responses”.⁷ A crawler hit is an access record. Treating crawl volume as an AI visibility metric reads a request log as an outcome.

Which AI platforms does Clarity cover?

Partially, and the answer is in the product rather than the documentation. The AI visibility report labels its own data source as “Microsoft Copilot and partners”. Microsoft’s published documentation is vaguer, referring throughout to “AI-generated answers” and “AI assistants” without naming anything, while the announcement covers “supported AI experiences” and notes that Microsoft “will expand coverage to additional partner AI platforms”.³ Check the label in your own dashboard rather than relying on a third-party account of the coverage, because it is the most specific statement Microsoft makes and it will change as that partner set expands.

Copilot is therefore confirmed, and the feature’s own marketing page reaches wider, claiming Clarity shows what is referenced “across not just Microsoft, but multiple AI platforms”.⁴ The same page is more precise elsewhere, describing the input as “Microsoft AI and select partner grounding data”.⁴

That phrasing is the most useful statement Microsoft makes, because it names the mechanism rather than a platform list. Microsoft can only observe grounding it serves. A platform that retrieves through Microsoft’s stack (the Bing index, or the Web IQ layer that ChatGPT uses for some of its web answers¹⁰) generates retrieval calls Microsoft handles and can therefore report on. A platform running its own index does not. So “multiple AI platforms” is compatible with “Copilot and partners”: the coverage extends past Microsoft’s own products, but only to platforms whose grounding runs through Microsoft.

What follows from that is firmer than any of the labels. Gemini, Perplexity, and Google AI Overviews retrieve through their own infrastructure and do not share grounding data with Microsoft, so they are out regardless of how the partner set grows. The partner list itself is unpublished, so the data should not be reported as ChatGPT coverage on the strength of the Web IQ connection alone.

The practical consequence is the same either way. A rising share of authority in Clarity is evidence of a trend on Copilot and whatever partner surfaces are currently included. Presenting the number to stakeholders as “AI visibility” without naming Copilot overstates what it covers.

How do you set up Clarity for SEO?

The Citation dashboard requires domain verification, not tag installation. A project administrator can verify by installing the Clarity tracking code, connecting through Google Search Console, or connecting through Bing Webmaster Tools.² The Search Console and Bing Webmaster Tools routes matter: they mean a site can use the citation data without adding another JavaScript tag, which removes the usual performance and consent objections. Verification is one-time per domain.

Bot Activity is separate and needs the CDN connection described above. The behaviour analytics side (recordings, heatmaps) does require the tag.

Where does Clarity fit alongside GSC and GA4?

It supplements rather than replaces. Google Search Console remains authoritative for Google impressions and clicks, including the Search Generative AI performance report. GA4 remains the on-site behaviour and conversion layer, and its AI Assistant channel group covers ChatGPT, Gemini, Claude, Copilot, and others by name, so it reports referral traffic from platforms Clarity’s citation data does not reach.

Clarity’s addition is the middle of the funnel that neither tool reports: which queries retrieved your content, which pages were cited, and which crawlers came for them. Read alongside AI visibility measurement tooling, it gives an observed baseline to check simulated citation rates against. If a prompt panel reports strong citation performance and Clarity shows no corresponding grounding activity, that gap is worth investigating before either number goes in a report.

Frequently asked questions

Is Microsoft Clarity free?
Yes, including the AI Visibility dashboards. The exception is Bot Activity, where the dashboard is free but your CDN provider bills you directly for log access, data transfer, and processing.⁷

Do I need the Clarity tag on my site to see citation data?
No. Domain verification through Google Search Console or Bing Webmaster Tools is sufficient for the Citation dashboard.² The tag is only required for behaviour analytics and for tag-based verification.

Does Clarity replace a paid AI visibility platform?
Not for competitive tracking. Clarity reports what happened to your domain on Copilot and partner surfaces; it cannot report on topics where you earn no citations, and it does not reach Gemini, Perplexity, or Google AI Overviews at all. For a defined competitive prompt set across named platforms, a dedicated tool still does something Clarity does not.

Does the citation data come from Clarity’s bot activity logs?
No, and the distinction matters for reading both dashboards. Microsoft lists the two as separate inputs, citing “Microsoft AI and select partner grounding data” and “integrated CDN-level data” as distinct rows.⁴ The behaviour confirms it: citation data works on Search Console or Bing Webmaster Tools verification alone, while Bot Activity stays switched off until you connect a CDN,⁹ so citations cannot be coming from your server logs. A server log also cannot contain a grounding query, because it records the URL requested rather than the retrieval string the model generated. Microsoft puts the split plainly, saying Bot Activity “shows who is accessing your content and how often” while Citations show “whether that access results in grounding, citation, or visible attribution”.⁹ Bot Activity reads your side of the request; Citations reads Microsoft’s.

Are grounding queries the same as search queries?
No. A grounding query is the retrieval string an AI system generated to find source content, which may differ substantially from what the user typed.² Treat them as a view into retrieval behaviour, not as keyword data.

Footnotes

1. Citations now generally available — Microsoft Clarity Blog ↩

2. Citation Dashboard Overview — Microsoft Clarity documentation ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

3. Understanding your influence in AI answers with Microsoft Clarity — Microsoft Clarity Blog ↩ ↩²

4. AI visibility — Microsoft Clarity ↩ ↩² ↩³ ↩⁴

5. Introducing Topic Insights: actionable content recommendations — Microsoft Clarity Blog ↩ ↩² ↩³

6. Analyse AI citation queries by topic — Microsoft Clarity Blog ↩

7. Bot Activity overview — Microsoft Clarity documentation ↩ ↩² ↩³ ↩⁴ ↩⁵

8. Troubleshooting AI Visibility — Microsoft Clarity documentation ↩

9. Clarity FAQ: AI Visibility — Microsoft Clarity documentation ↩ ↩² ↩³

10. Microsoft Web IQ, with Jordi Ribas, President of Search and AI, quoted in Microsoft releases Web IQ, powered by Bing but designed for how AI-agents search — Search Engine Land ↩