What "AI visibility" means, and why it's slippery
PPC Land's explainer defines AI visibility as how often a brand, product or website is named, cited or recommended in answers from systems such as ChatGPT, Gemini, Claude, Perplexity, Microsoft Copilot and Google's AI Overviews. The label belongs to vendors, not to a standards body.
The nearest thing to a shared vocabulary is guidance the Interactive Advertising Bureau published on August 3, 2026. The IAB's announcement says more than 20 companies sell AI visibility measurement tools. Those tools use different methods and can give different answers for the same brand or publisher.
The IAB framework groups metrics into four levels, the "4 P's":
- Presence: mention rate, citation rate, share of voice and visibility momentum.
- Prominence: placement or order, and how substantively content is used.
- Portrayal: sentiment, framing, hallucination rate and factual inaccuracy rate.
- Persuasion: recommendation strength and post-citation click-through rate.
The IAB also splits data into two tiers. Directional data is good for spotting trends. Decision-grade data needs more rigor in query volume, sampling, prompt coverage, cadence, reproducibility and platform coverage. The IAB says only the second tier should drive budget or strategy decisions. This is vocabulary and disclosure guidance, not an audit or certification, and it doesn't make vendor scores interchangeable.
Why the statistics are contested
AI answers are non-deterministic, so the same prompt can return different brands on different runs. PPC Land reports a SparkToro study, with Gumshoe.ai, in which 600 volunteers ran 12 prompts 2,961 times. According to that account, the chance of two identical brand lists was under 1 in 100. Rand Fishkin argued that a "ranking position in AI" is meaningless but that visibility percentages over many repeated prompts are reasonable. PPC Land notes that the work was not peer-reviewed and was done with a partner whose business is visibility tracking.
I haven't independently verified the SparkToro figures. Treat them as reported by PPC Land.
Third-party trackers have a second problem: the vendor chooses the prompt library. The IAB asks providers to disclose library size, refresh cadence and whether queries are synthetic.
The Microsoft angle: first-party citation data
Third-party tools guess at what users see. Microsoft's own reports show citations from its own systems for verified sites.
Bing Webmaster Tools: AI Performance
Microsoft introduced the public preview in February 2026. It is a free report that shows when pages from a verified website are cited in supported AI-generated answers. The surfaces covered are Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations.
The launch metrics were:
- Total citations: how often your content was displayed as a source in the selected period.
- Average cited pages: the average number of unique pages shown as sources per day.
- Grounding queries: the key phrases the AI used when retrieving content that was later referenced. Microsoft says this is a sample of overall citation activity.
- Page-level citation activity and a timeline of changes.
Microsoft's launch post is explicit about what these numbers are not. They do not indicate ranking, authority, page importance, or placement within an individual answer. Bing says it respects robots.txt and other supported content-owner controls.
Grounding queries are the part most often misread. They are not what a user typed. They are Copilot's own reformulation of the question when it went looking for sources. That makes them useful for seeing how Microsoft's systems categorise your content. They are not a literal keyword report.
A later addition connects the Grounding Queries and Pages views. You can select a query to see its cited pages, or a page to see its queries. One query may map to many pages and vice versa.
The June 2026 expansion
On June 16, 2026, Microsoft announced four preview capabilities, rolling out globally:
- Intents: grounding queries are classified into categories such as Informational, Commercial, Navigational, Learn and Solve, Research, Creation and Local.
- Topics: related grounding queries are grouped into thematic clusters. Microsoft's example is "solar panels" and "residential solar installation" landing under a broader solar energy topic.
- Citation Share: the percentage of citations attributed to your site out of all citations shown for the same grounding query.
- Compare: overlays a previous period, such as the prior 30 days, onto the current view.
Microsoft calls Citation Share an observational metric, not a ranking system or competitive scoreboard. It does not expose competitor domains, represent traffic share, or score content quality. Microsoft also says the Intents and Topics classifiers are still maturing. During preview, some labels may be broad, especially for niche domains.
Microsoft Clarity: Citations and share of authority
Microsoft Learn documents a Citations dashboard under AI Visibility in Clarity. Its key metrics are:
- Page citations: how many times your pages were referenced in AI-generated answers during the period.
- Share of authority: your citations divided by total citations from all domains.
- Branded and non-branded share of authority: the same calculation split by query type. It shows whether visibility comes from brand recognition or from topical coverage.
- AI referral traffic: the percentage of sessions that arrived from AI assistants.
- Query topics, grounding queries and My cited pages: supporting views.
The share-of-authority method has a quirk. The calculation is daily. If your domain is cited for a query that day, all citation instances for that query that day count. If it isn't, that day is excluded. Microsoft's documentation says this focuses on active participation and can produce higher values than broader calculations. A "good" number may partly reflect the method.
Setup for Clarity
Clarity's overview says a project administrator must verify domain ownership before citation data appears. The documented options are:
- Install the Clarity tracking code on the site.
- Connect the domain through Google Search Console.
- Connect the domain through Bing Webmaster Tools.
With the Search Console or Bing option, you are redirected to sign in and select the domain. Verification then typically completes automatically. It is a one-time step per domain. After that, choose AI Visibility, then Citation, from the dropdown menu.
Clarity is also explicit that the dashboard tracks citations only. It does not measure traditional rankings, impressions or click-through rates.
What the data can and can't tell you
These reports can answer bounded questions:
- Which of your pages Microsoft's AI experiences cite.
- Which grouped phrases are associated with those citations.
- Whether citation activity is rising or falling over time.
- How your share compares with the full set of cited sources, per Microsoft's definitions.
They cannot show:
- That a specific edit caused a change. Microsoft says citation activity can shift with evolving models, competing content, freshness signals and demand, and Compare may only reveal shifts that "correlate" with those factors.
- Your presence in ChatGPT, Gemini, Claude or Google's AI features. PPC Land notes that none of the first-party reports covers rival platforms.
- Visits or business impact. A citation is not a click. Citation totals can be high while referral traffic stays flat.
Microsoft's own advice is the same as the IAB's. Treat the numbers as directional monitoring, not guaranteed accuracy. PPC Land attributes that framing to the July 2026 Clarity Topic Insights release, and I haven't checked that page myself.
A practical workflow
- Verify your site in Bing Webmaster Tools and open AI Performance.
- Start with page-level citations to see which URLs are used as sources.
- Use the query-to-page mapping to see which grounding phrases pull each page in.
- Use Compare on 30-day windows. Look for trends over weeks, not day-to-day spikes.
- Cross-check against analytics. If citations rise and sessions don't, that is a known pattern, not a tracking error.
- Use Microsoft's suggested content fixes sensibly: clear structure, supported claims, fresh content, and consistent text, images and video. Microsoft also points to IndexNow to notify participating search engines when content changes.
- If you run a third-party tracker as well, check its disclosed library size, cadence and sampling against the IAB's criteria.
Bottom line
AI visibility has no audited standard yet. The IAB's framework gives buyers better questions to ask. Microsoft's Bing and Clarity reports give site owners ground truth for one slice of the AI web: Copilot, Bing summaries and select partners. Treat that slice as an indicator, not a scoreboard. The reports are useful, but they won't tell you how you rank in an AI answer.