Microsoft Clarity AI Citations: What It Tracks (and What It Misses)
Clarity's Citation dashboard is free, first-party data on which of your pages AI answers cite. Here's what its metrics mean, what the new scrape-to-referral ratio tells you, and the one ecosystem it can see.
Updated · Published
Frequently asked questions
Does Microsoft Clarity track AI citations?
Yes. Clarity's Citation dashboard, part of its AI Visibility reports, shows which of your pages were referenced in AI-generated answers, how often, and which queries retrieved them. It became generally available on 13 May 2026. Microsoft's documentation is explicit that it tracks citation activity in AI answers only, not search rankings, impressions, or click-through rates.
Is Microsoft Clarity's AI Visibility free?
Yes. Microsoft's own FAQ states that "Clarity is a free service forever" with no traffic limits and no paid upgrade, and the AI Visibility reports are part of the product. The documented caveat is eligibility rather than price: AI Visibility features are available for eligible website categories, so some projects may not have access to every report.
Does Clarity show ChatGPT or Perplexity citations?
No, not directly. Microsoft's documentation says Citations covers "supported AI experiences" without naming them, and the data is built from Microsoft's own grounding and crawl signals. ChatGPT's own search, Gemini, and Perplexity do not share their internal grounding logs with Microsoft, so citations there do not appear in the dashboard.
What is a grounding query in Microsoft Clarity?
A grounding query is the search phrase an AI system used to retrieve your content before writing its answer. Microsoft notes these "may differ from the exact phrases users type," because the model rewrites a question into retrieval terms. The table reports each query alongside volume, citation frequency, and citation rate.
What is share of authority in Clarity?
Share of authority is the percentage of total citations attributed to your domain: your citations divided by total citations from all domains, calculated daily. Microsoft notes that days when your domain was not cited for a query are excluded, so the metric "focuses on active participation and may lead to higher Share of authority values compared to broader calculations."
Does Microsoft Clarity separate branded and non-branded AI citations?
Yes, since 3 August 2026. Grounding queries carry a branded or non-branded label, share of authority is broken out along the same split, and the whole dashboard can be filtered to either segment. Microsoft's stated purpose is to show whether performance "reflects direct brand visibility, category discovery, or a mix of both." The non-branded segment is the competitive read; branded share of authority largely reflects demand you generated elsewhere.
Is Clarity a replacement for an AI visibility tool?
Not on its own. Clarity gives you observed citations from one ecosystem's logs, which is data no prompt-based tool can reproduce. It cannot tell you what an answer said about your brand, which competitors were named beside you, or what happens in ChatGPT, Gemini, and Perplexity. Most teams need both the log view and cross-engine prompt sampling.
Does Clarity report AI referral traffic?
Yes. The AI referral traffic metric shows the percentage of sessions on your site that arrived from AI assistants, calculated as AI-referred sessions divided by total sessions for the period. That makes it a rare free source for the downstream number Google's Search Console AI report omits, since that report launched with impressions but no clicks or click-through rate.
Does Microsoft Clarity show whether AI crawlers send traffic back?
Yes, since 13 August 2026. The AI Scrape-to-Referral Ratio card in the Bot Analytics dashboard compares AI scrape activity against the referral traffic each operator sends back, ranks operators by it, and links through to session recordings filtered to the selected AI source. It requires connecting a supported CDN — Cloudflare, Fastly, or Amazon CloudFront, or the WordPress plugin — because the scrape side comes from server logs, not Clarity's JavaScript tag.
Microsoft Clarity now reports which of your pages AI answers cite, for free, using data you cannot get from prompt tracking. The Citation dashboard went generally available on 13 May 2026, and it is built from the citation logs of the systems that generated the answers rather than from sampling. That makes it genuinely new. It also sees one ecosystem, which is the part most coverage skips.
Key takeaways
Clarity is free with no traffic limits, and the Citation dashboard is part of it. The gating factor is eligibility, not price: AI Visibility is "available for eligible website categories."
It is observed citation data, not sampled. Microsoft is reporting what its own retrieval actually did, so there is no prompt set to design and no run-to-run noise.
Coverage is the catch. Microsoft's docs say "supported AI experiences" without naming them, and ChatGPT's own search, Gemini, and Perplexity do not hand their grounding logs to Microsoft.
Grounding queries are the most useful part: the phrasings the engine actually retrieved on, which are not the phrasings users typed.
Read share of authority carefully. Microsoft says its daily calculation "may lead to higher Share of authority values compared to broader calculations."
Since 3 August 2026 that metric splits by branded and non-branded grounding queries. Only the non-branded half is a read on your content rather than on demand you already had.
It reports AI referral traffic, the downstream number Google's own AI report still does not give you.
Since 13 August 2026, a scrape-to-referral ratio in Bot Analytics puts a per-operator number on whether AI crawling is buying you any visits — if you can meet the CDN requirement that feeds it.
Clarity's Citation dashboard sits under AI Visibility and reports five things. Microsoft's definitions are precise enough to be worth taking literally.
Metric
What Microsoft says it is
Page citations
"The total number of times pages from your site were referenced in AI-generated answers during the selected period."
Share of authority
"The percentage of total citations attributed to your domain, calculated as your citations divided by total citations from all domains."
AI referral traffic
"The percentage of sessions on your site that arrived from AI assistants."
Grounding queries
"The queries that AI systems used to retrieve your content before generating an answer."
My cited pages
The pages referenced in AI answers, "along with their citation counts and associated queries."
Two notes in the documentation do more work than the metric definitions. The first sets the scope: "The Citation dashboard tracks citation activity in AI-generated answers only. It does not measure traditional search rankings, impressions, or click-through rates." The second is a warning against reading citations as placement: counts "reflect how often a page was referenced, not its ranking or prominence within an AI-generated answer." Being cited third in a hedged paragraph and being the source an answer is built on count the same.
Setup is a one-time domain verification by a project administrator, through the Clarity tracking code, Google Search Console, or Bing Webmaster Tools. On 9 July 2026 Microsoft added Topic Insights on top, grouping citations into topics with content recommendations. It ships in beta, and Microsoft's own framing is that it "should only be used for directional monitoring and insights." Three more releases followed within five weeks: query topics on 22 July, which group raw grounding queries into generated themes, the branded/non-branded split on 3 August, and the scrape-to-referral ratio on 13 August — both covered below.
Here is the distinction that makes this worth your attention. There are two ways to know whether AI answers cite you, and they answer different questions.
Observed data comes from the logs of the system that generated the answer. Clarity can tell you your pages were cited 400 times last week because Microsoft's infrastructure did the citing. There is no sample, no estimate, and no prompt set to argue about.
Sampled data comes from running prompts yourself and reading what comes back, which is the method behind prompt tracking and share of voice. It is an estimate, and it is engine-agnostic: you can point it at anything with an API or an interface.
Clarity (observed)
Prompt tracking (sampled)
Source of truth
The platform's own citation logs
Answers you generate and record
Engine coverage
"Supported AI experiences" only
Any engine you choose to query
Tells you the answer's wording
No
Yes
Tells you who was named beside you
Only as competing domains
Yes, by brand
Sentiment and accuracy
No
Yes
Cost
Free
Your time, or a tool, plus API usage
Neither supersedes the other. Observed data settles volume questions that sampling can only approximate. Sampling answers the questions a log cannot: what the engine said about you, whether it got your pricing right, and which three competitors it recommended instead. Our study of citation volatility is an example of the second kind, and it also explains why a single sampled check is weak evidence: across 28 prompts tracked daily for six weeks, the set of cited domains turned over by roughly 60–70% day to day.
Microsoft describes Citations as surfacing activity "across supported AI experiences" and does not publish the list. What the product does disclose is where the data comes from: it unifies signals from Bing Webmaster Tools and Clarity, spanning crawling and indexing through to citation and AI-referred traffic. That is Microsoft's own retrieval stack.
The practical read, as Search Engine Journal put it in May 2026, is that Clarity captures citations "primarily across Microsoft's AI surfaces (like Copilot and Bing generative search)" and "is not giving you a direct window into how OpenAI's ChatGPT (using its own search), Google Gemini, or Perplexity are citing your links, because those platforms do not share their internal grounding logs with Microsoft."
That is not a flaw in the tool; it is the structural condition of first-party AI data. Every platform can only report on its own retrieval, and none of them will report on each other's. So a dashboard fed by one ecosystem's logs is precise and narrow at the same time. Treat the number as accurate for what it covers, and do not let it stand in for AI citations generally.
Share of authority looks like a share-of-voice metric and behaves slightly differently, in a way Microsoft documents openly. The calculation runs at a daily level: "If your domain is cited for a query on a given day, all citation instances for that query on that day are included; if not, that day is excluded." Microsoft then says what follows from that, which is the useful part: the approach "focuses on active participation and may lead to higher Share of authority values compared to broader calculations."
In other words, days you lost are not fully in the denominator. A 30% share of authority is not the same claim as "we appear in 30% of relevant AI answers." It measures your slice of the citations on the queries and days where you were already in the running, which is a narrower and more flattering base.
That makes it a good trend line and a bad boardroom number. Watch it move, and pair it with a presence-style measure across your actual prompt set if you need to answer "how often do we show up at all?"
On 3 August 2026 Clarity started labelling each grounding query as branded or non-branded, breaking share of authority out along the same line, and letting you filter the dashboard to either segment. Microsoft's stated reason is the ambiguity in the blended figure: without separating the signals, "it can be difficult to understand whether your performance reflects direct brand visibility, category discovery, or a mix of both."
That is a partial fix for the problem in the section above, and a more interesting one than a filter usually is. A branded grounding query names your company, and on questions about your company you are the obvious source — the engine was always likely to ground on your own pages. Your share of authority there runs toward a ceiling that reflects demand generated elsewhere, by your marketing, your press, or your category position. Non-branded grounding queries are the ones where the engine had a real choice among domains, and where your competitors sat in the same denominator.
Segment
What a high share of authority tells you
Branded
AI answers about you are grounded in your pages rather than a review site's summary of them. Worth confirming, but not a result your content earned.
Non-branded
Your page won a citation on a topic anyone could have supplied. This is the competitive number.
Two limits are worth carrying. Microsoft has not published how the classification works. PPC Land, reading the announcement's sample table, notes that the brand term appears highlighted inside the query text, which suggests substring matching — and either way, nothing in the announcement addresses misspellings, product names that differ from the company name, or comparison queries containing a competitor's brand alongside yours. And the feature was announced on the Clarity blog rather than specified in the docs — as of this update the Citation dashboard reference still describes share of authority as a single figure — so the guidance available is a launch post, not documentation.
The split is also a template worth copying into whatever you run outside Microsoft's stack. If your tracked prompt set is mostly your brand name plus a modifier, you are measuring that same ceiling and calling it AI visibility. The prompts that tell you something are the category questions a buyer asks before they know you exist, which is the distinction behind AI competitor analysis and the reason prompt sets are worth designing rather than accumulating.
On 13 August 2026 Clarity added an AI Scrape-to-Referral Ratio card to its Bot Analytics dashboard, comparing AI scrape activity against the referral traffic each operator sends back, so that — in Microsoft's framing — you can "evaluate the tradeoff between content extraction and traffic return." It ships with a ranked list of operators by referral performance and a click-through into the Recordings tab pre-filtered to the selected AI source, so the question after "did they send anyone" — "did those visitors do anything" — is one click away.
This card answers a question the Citation dashboard never touched. Citations tell you AI answers reference your pages; the ratio tells you what the extraction behind those answers costs and returns. It is the site-level version of the metric Cloudflare introduced on Radar in July 2025, whose network-wide numbers set expectations: in Cloudflare's launch-week data, Anthropic's crawlers fetched roughly 70,900 HTML pages for every referred visit. The sample dashboard in Microsoft's own announcement shows about 6,000 scrapes per referral. If your ratio comes back in the thousands, that is the ordinary reading, not an anomaly.
Three caveats before you act on it.
The scrape side needs a CDN you may not have. Referral sessions come from Clarity's JavaScript tag, but crawler requests never execute JavaScript, so the bot side is built from server logs delivered through a connected CDN — Cloudflare, Fastly, or Amazon CloudFront, with WordPress sites covered automatically via the Clarity plugin. Microsoft also notes that log delivery may incur charges billed by your CDN provider. Without one of those integrations, the card has nothing to divide. Microsoft additionally restricts the calculation to correctly mapped domains, flagging when bot data and referral data cover different ones.
The denominator is undercounted. AI referral sessions are only attributable when a referrer survives the click, and visits from in-app browsers or noreferrer links arrive as direct traffic — the same limitation that caps every AI referral measurement. An undercounted denominator inflates the ratio, so read it as an upper bound on the imbalance.
The numerator mixes motives. Scrape counts include training crawls that were never going to produce a referral, alongside retrieval fetches that might. A brutal ratio is an argument for revisiting which crawlers your robots.txt allows, but the training/retrieval split — which that decision actually turns on — is not in the card.
The decision rule that survives all three: use the operator ranking and the session recordings to compare your own bots against each other and over time. Do not benchmark your ratio against Cloudflare's network figures or anyone else's Clarity screenshot — the inputs differ enough that cross-site comparison is noise wearing a decimal point.
The metric with the most to teach you is the least headline-friendly. Grounding queries are the phrasings the engine used to retrieve, and Microsoft is upfront that they "may differ from the exact phrases users type." A model takes a conversational question, decomposes it, and issues its own retrieval terms, which is the mechanism that makes grounding work at all.
Seeing those terms is close to reading the engine's notes. You learn which of your pages satisfied which retrieval intent, where a single page is absorbing queries that deserve their own coverage, and which adjacent phrasings retrieve competitors instead of you. That is a concrete content brief, and it is the kind of input answer engine optimization usually has to infer.
The caveat is the same one as above. These are one ecosystem's retrieval patterns. They generalize reasonably, because engines that ground answers face similar retrieval problems, but "generalizes reasonably" is an assumption, not a measurement.
If you are assembling AI visibility measurement without a budget, Clarity is a real addition rather than another grader. It gives you observed citations, grounding queries, AI referral traffic, and — with a CDN connected — a per-operator scrape-to-referral ratio for nothing, which no free prompt-based option can match on precision. Expect the referral percentage to be small in absolute terms — AI traffic is a fraction of a percent of visits on most sites, however well it converts — so read it as a trend, not a channel.
It leaves three gaps: the engines outside Microsoft's stack, the wording and sentiment of the answers themselves, and the competitor set named beside you. Those need sampling. The method is in our guide to tracking your brand in AI search, and Elmo automates it as an open-source, self-hosted platform that runs your prompts across ChatGPT, Claude, Gemini, Perplexity, and Google's AI surfaces and records every mention, citation, and rival named alongside you. For the rest of the no-cost options, see our roundup of free AI visibility tools; for the paid category, the best AI visibility tools comparison and the AI visibility software hub.
The short version: turn Clarity on, because free observed data about your own citations is worth having. Just do not mistake one ecosystem's logs for a view of AI search.