Microsoft today opened the brand-matching list behind Clarity's AI Visibility dashboard to manual revision, allowing site owners to add abbreviations, translations and typos to the terms that decide whether an AI query is classified as branded. According to Microsoft, the initial list is pre-populated by Clarity and can be changed under Settings > AI Visibility.

In Short

Microsoft Clarity, a free web analytics service, now allows site owners to change the list of names and spellings it uses to decide whether an AI query is about their brand. A brand often turns up in AI queries as an abbreviation, a translation or a misspelling, so a list that misses those forms can file a query in the wrong group and skew the branded and non-branded figures. The list still starts out filled in by Clarity, but it can now be adjusted, with related spellings grouped under a single brand name.

What Microsoft changed in the AI Visibility settings

According to Microsoft, site owners can now customize the Brand Terms that Clarity uses to classify branded queries in the AI Visibility dashboard. The post appeared on the Clarity blog today under a Clarity Staff byline, and it follows the branded and non-branded query segmentation that Microsoft added to the citation dashboard on August 3, 57 days earlier.

Microsoft describes the change in four parts. The first is management: Brand Terms can be added, changed or removed from Settings > AI Visibility, so that the list reflects how brands, products and the business are referenced. Clarity pre-populates the initial list, which Microsoft presents as a starting point for refinement rather than a finished set.

The second concerns variation. Terms can be added for abbreviations, alternate spellings, localised names, translations and common typos. Microsoft's own example uses its name: a Brand Name such as Microsoft could carry the related Brand Terms Microsoft, MS and Microsoft Corporation.

The third is structure. A Brand Name acts as the parent entry, and related Brand Terms sit beneath it. When analysing AI visibility, the dashboard can be filtered by Brand Name, which returns results across every associated term, or by one individual Brand Term to isolate how a single variant performs.

The fourth is upkeep. Keeping names and terms current, according to Microsoft, helps Clarity recognise the queries tied to a brand, "improving classification accuracy and reporting quality."

What the illustration shows

The image published with the post shows two figure cards, Branded SoA at 53.83% and Non-branded SoA at 16.24%, with a pencil icon and a tooltip on the first card. Beneath them sits a panel with a button labelled New brand name and a two-column table headed Brand name and Alias. Two rows are visible. Alpine ski carries the variants Alpine, Alpine sik and Alpien ski; Alpine ski house carries Alpine ski home and Alpin ski house. Several entries read as keyboard slips or near-spellings - a dropped letter, a transposition, a shortened form of Alpine - which matches the typo category the text lists.

One terminology gap shows in the picture. The table column is headed Alias, whereas the text calls the same entries Brand Terms, and the post does not say whether the two words describe one field.

The gap between the two cards is 37.59 percentage points, or roughly 3.3 times. The post does not expand the abbreviation SoA. Microsoft's August release split a Share of Authority card by query type, so SoA most likely refers to that metric. That is a reading, not a statement in the new post, which also leaves open whether the figures come from a live project or were composed for the picture.

What the post leaves open

The announcement is short, and several operational details that govern how the feature behaves are missing.

Matching and reclassification

The post does not say how Clarity compares a query with a term: exact match, substring, or some tolerance for misspellings. When Microsoft added the branded split, its sample table highlighted the string alpine inside branded queries, which pointed to substring matching that was never documented. The typo entries in the new illustration, Alpine sik and Alpien ski, suggest that variants are matched as listed strings and not detected automatically, although that too is an inference from the picture.

Nor is it stated whether adding a term reclassifies queries already in the dashboard or applies only to data collected afterwards, or how quickly a change takes effect. For a metric followed over time, the difference decides whether a trend line stays comparable across the moment a list is revised.

Limits, scope and cost

No ceiling on Brand Names or Brand Terms is given. In paid search, Microsoft Advertising caps brand lists at 20 per campaign in AI Max; the lists serve a different purpose, but limits of that kind exist elsewhere in the company's tooling. Product names are covered, since the settings text refers to brands, products and the business. Competitor names that appear in comparison queries go unmentioned, as do multi-word brands that contain common nouns - two cases that the August release also left undefined. Misspellings and product names, by contrast, now fall to the site owner to list.

Microsoft has noted that AI Visibility features are available only for eligible website categories, and today's post does not revisit that. It carries no pricing note either. The page footer describes Clarity as free, with no limits on traffic, and earlier AI Visibility releases shipped at no additional cost.

Where the change sits in Clarity's AI Visibility sequence

Read against the earlier releases, the change alters who holds the classification input. Query Topics sorts groundingqueries into generated themes automatically, and Microsoft has not published the classification methodology. Topic Insights, by contrast, works from topics that users define. The August branded split gave no account of how brands were identified; the new post at least indicates that a Clarity-generated list sits underneath. Today's release moves that boundary into the site owner's hands while leaving the default in place.

The cadence has been quick: four AI visibility additions in roughly five weeks by mid-August. Clarity's blog lists a post on page classification in Bot Activity as the entry immediately before today's, though the source carries no date for it.

Why the classification list matters for measurement

The split between branded and non-branded demand is among the oldest divisions in search marketing, and it has already moved into AI-era tooling. Google's Search Console offers a branded and non-branded query filter, and its guidance says branded queries include common misspellings. Google Ads added a native branded search toggle inside AI Max on June 1. On the traffic side, Similarweb's study of downstream visits left out users who named the brand in their prompt, so that prior awareness would not contaminate the result.

A term-based classification can err in two directions. A missed variant pushes a branded query into the non-branded group and lowers the branded share; an over-broad term does the reverse. Two projects with identical citation records but different lists would therefore report different splits. Does a list that a site owner writes measure the same thing as one Clarity wrote? The answer depends on how the matching works, which the post does not describe.

The industry has not settled its definitions. An IAB framework published on August 3 found that only 16% of brands systematically track AI visibility, and that more than 20 companies sell measurement tools built on different methodologies that produce different answers for the same brand or publisher. A user-adjustable term list is one input to that spread which Clarity now exposes. PPC Land's analysis of the August release read the branded share as something close to a recognition index measured at the retrieval layer, and the non-branded share as the figure showing whether content does work an AI system could not do without it. Both readings depend on where the line between the two is drawn.

What remains to be seen is documentation. Whether Microsoft publishes the matching rules and the treatment of historical data will determine how comparable branded shares are between sites that configure their lists differently.

Timeline

Summary

Who: Microsoft, through the Clarity analytics team, which published the post on the Clarity blog under a Clarity Staff byline. The change concerns site owners, brands, publishers and marketers who use the AI Visibility dashboard.

What: Customizable Brand Terms in the AI Visibility dashboard. Terms can be added, changed or removed under Settings > AI Visibility, grouped beneath a Brand Name, and used to filter results by Brand Name or by individual term. Clarity pre-populates the initial list.

When: Today, September 29, 2026, 57 days after the branded and non-branded query segmentation of August 3, 2026.

Where: Inside Microsoft Clarity's AI Visibility dashboard and its Settings > AI Visibility page.

Why: According to Microsoft, brands appear in AI queries as abbreviations, alternate spellings, localized names, translations and typos, and customizing the terms can improve the accuracy of branded versus non-branded segmentation and reporting.