Skip to content
Today
Sign in

BuildNot yet confirmed elsewhere1 publisher2 min readPublished

Google AI Overviews climb from 26% to 90% of DemandSphere's tracked branded searches in September

Google showed AI Overviews on 90.48% of DemandSphere's tracked branded keywords on September 27, up from 26.12% on September 1. The count records only whether a summary appears, so teams must check its wording and citations before reading anything into traffic.

The Engineer · Build desk

How we use AISend a correction

Photograph accompanying Google AI Overviews climb from 26% to 90% of DemandSphere's tracked branded searches in September
Photo: demandsphere.com

What happened

  • Search Engine Land reported the same finding on October 1 and noted that Google had not publicly announced any change behind the increase.
  • The share eased after the peak but stayed high, at 80.23% on September 28 and 82.06% on September 29.
  • The branded set spans company, product and service names plus navigational searches, queries the source describes as often high-intent.
  • An Overview can appear on a brand's query without citing the brand at all, or can answer it from third-party sources.

Compiled by The EngineerSomething wrong?How this is made

Why it matters

  • constraint A position-one organic ranking for a brand's own page leaves the generated layer unmeasured on most of the panel's brand queries, so rank reports alone undercount what searchers are shown.
  • exposure By the source's argument, third-party pages cited in an Overview can shape a high-intent brand search before the searcher reaches the company's own site.
  • decision Any claim of lost traffic from this curve has to wait for a join against the brand's own click data, since the panel does not measure clicks, conversions or citations.

DemandSphere's counter is a yes-or-no check. For each tracked branded keyword, every day, across markets and devices, it records whether Google returned an AI-generated summary [4][5]. Whether that summary cites the brand being searched does not change the count [5]. An Overview that credits a competitor scores the same as one that links the brand's own help page [5].

Daily sampling across markets and devices is the right design for watching a feature spread, and the source credits the method for exactly that [15]. The curve it caught is steep. The share rose 64.36 percentage points in 26 days, about 3.5 times the September 1 level [17][18][21]. On September 1 roughly three in four tracked keywords returned no Overview. At the peak, fewer than one in ten did [19]. The fall on September 28 was 10.25 points in a single day [20].

That overnight swing is why I would not plan around the peak. The 90.48% describes DemandSphere's keyword set [2]. For it to describe a given brand, that brand's priority queries would have to resemble the panel in market, device and query mix. The source does not say how many keywords the panel holds or how they were chosen. In an internal panel I would split queries by type, so a rise in one group cannot hide inside the average.

The cause is unconfirmed. The source warns against reading the pattern as a specific ranking update, a new citation policy, or a change to every branded query [9].

The part a team can reuse is the observation. DemandSphere's method does not establish whether a company was named, linked, recommended or accurately described in an answer [8]. The source's checklist fills that gap with per-query records: whether an Overview appears for priority company, product and service queries, the wording used about the brand, the sources cited, and whether the company's own pages are among them, compared over time, markets and devices [11]. A rank tracker stores an organic position per keyword [12]. Most of these new fields are free text. The source says deciding whether a generated description is correct, or whether its cited sources raise a competitive concern, is an interpretive call that automated monitoring does not make [16].

What to watch

  • A public explanation from Google for the September rise, or confirmation of a rollout on branded queries.
  • DemandSphere's October readings: whether the share holds near the 80-82% of late September or falls back toward 26%.
  • Brand-side data joining Overview observations to clicks, the first evidence on traffic that this panel cannot supply.

Clarity's read

What the record supports and how the coverage leans. The claims behind it follow.

Reality

Evidence45
Adoption60
Hype gap+5
Incentives65
Confidence45
Why these scores

Claim ledger

Ranked by verification strength, evidence, and original report placement.

  1. [1]

    DemandSphere recorded AI Overviews on 26.12% of tracked branded keywords on September 1, 2026.

    ReportedSupportedSource: DemandSphere data, as reported on dev.toView cited source
  2. [2]

    AI Overview appearances rose to a monthly high of 90.48% of DemandSphere's tracked branded keywords on September 27, 2026.

    ReportedSupportedSource: DemandSphere data, as reported on dev.toView cited source
  3. [3]

    After the peak the level eased but remained high, at 80.23% on September 28 and 82.06% on September 29.

    ReportedSupportedSource: DemandSphere data, as reported on dev.toView cited source

Sources

1 independent publisher whose own reporting we read for this story.

  1. dev.to

    1 article · October 9, 2026

    Google AI Overviews Expand Across Branded Queries, DemandSphere Data Shows

Share your take

Let Clarity write the post for you.

Signed-in readers get a short post drafted on this story in the register they choose — narrative, analytical, or a direct position — editable to the last word before it goes anywhere. The share buttons at the top of this story work without an account.

Create a free account to composeSign in

Topics and entities

Follow any of these and your For You feed starts watching them — no settings page required.

Topics

Entities

Loading related stories