Google AI Overviews are taking over branded search

Branded search is no longer a safe click path. In late September 2026, Google AI Overviews surged across branded queries, with DemandSphere's daily tracking rising from 26.12% on September 1 to 80.23% on September 28, after peaking at 90.48% on September 27. In a separate test, Chris Long, co-founder of Nectiv, found AI Overviews on 93 of 100 brand searches. That changes the job for SEO, content, and brand teams. When someone searches your company by name, Google is increasingly willing to answer first and send the click second.

• Branded Google queries saw AI Overview presence jump from 26.12% to 80.23% in late September 2026.
• A separate test found AI Overviews on 93 out of 100 brand searches.
• Ranking first for your own name is no longer enough if Google summarizes your brand above or around the click.
• Teams now need to track overview presence, source selection, and brand framing, not just organic position.

Why does this jump matter so much?

Because branded queries were supposed to be the safest part of search. If a person types your company name, the expected outcome is simple: they find your site, your pricing, your login, or your product pages. When AI Overviews appear on those searches, Google inserts a machine-written interpretation between the user and the brand.

That is a bigger shift than it sounds. A branded query is usually high intent. The searcher already knows the company. They are not asking Google to introduce a category. They are asking to reach a specific brand. If Google now answers before the click, it controls the first summary the searcher sees, even when demand already exists.

The timing also matters. This was not a slow crawl upward across the whole month. DemandSphere's curve stayed mostly in the high 20s and low 30s for much of September, then accelerated sharply in the final stretch. That pattern suggests a material rollout or behavior change, even though Google has not publicly announced one.

There is also a practical nuance here. Most observed examples were not giant AI boxes permanently pinned above the official site. Many appeared mid-page or lower in the results. That makes the change less dramatic than a full replacement of the brand result, but it does not make it harmless. Mid-page still shapes the branded experience, especially on mobile or when the answer is expanded.

What did the September data actually measure?

It measured presence, not quality. DemandSphere tracked branded keywords once per day across all markets and devices and counted an AI Overview whenever Google returned one for the query. That means the spike from 26.12% on September 1 to 80.23% on September 28 tells us how often the AI layer appeared, not whether the brand was cited, described correctly, or positioned favorably.

That methodological detail matters. A brand can trigger an overview and still lose the narrative. Google's summary might rely on the brand's own pages, but it might also lean on third-party articles, directories, public knowledge panels, or older web pages that were never meant to become the opening sentence of the brand story.

The second data point strengthens the signal. Chris Long tested 100 enterprise brands through SerpApi and found AI Overviews on 93 of them. The observed examples included names like Reddit, Salesforce, Amazon, Adobe, Mailchimp, and Facebook. In those observations, Google's own brand name and news publishers were notable exceptions.

There was also variation in placement. Adobe reportedly showed an AI Overview at the top of the results, while many other brands showed the answer lower on the page. That is another reminder not to flatten everything into one metric. An overview that appears, an overview that appears above the homepage, and an overview that cites the brand's own pages are three different outcomes.

So the headline is real, but it has limits. This dataset is excellent for spotting a shift in Google's behavior. It is not enough on its own to tell a brand whether the shift is helping, hurting, or simply changing where attention goes.

Why is branded search now a GEO problem, not just an SEO problem?

Because the fight is no longer only about rank. Generative Engine Optimization (GEO) is the work of making sure AI systems can understand, trust, and reuse your brand correctly. Once Google starts summarizing brand queries at scale, the question changes from “Do I rank first?” to “What is Google saying about me, and where did it get that idea?”

This is where many reporting stacks break. Traditional SEO tools are built to tell you whether your homepage ranks, whether a knowledge panel appears, and whether a competitor moved above you. Those are still useful signals. They are no longer enough when Google's interface adds a generated answer that can synthesize multiple sources into a new layer of meaning.

Take the simplest example. If someone searches a brand like Adobe or Reddit, the old win condition was obvious: own the top organic result and make the snippet strong. The new win condition is messier. You still want the top organic result, but you also need Google's summary to be accurate, current, and grounded in the pages you actually want reused.

That is why branded visibility has to be treated as a wider AI search visibility problem. The brand homepage, help center, product pages, documentation, and third-party web footprint can all influence the generated layer. We made a similar point in why AI search traffic does not follow organic search rules: a page can keep its classic search position and still lose visibility when AI systems decide to answer in their own words.

That is also why branded search should not be dismissed as “just navigational.” In a search environment with generated answers, navigational intent still exists, but the interface no longer behaves like a pure handoff to the official site. Google now has room to interpret, summarize, and frame.

What should teams measure now besides rankings?

They should measure the generated layer directly. If branded AI Overviews are now common, the reporting model needs to expand beyond rank one, impressions, and branded CTR. The goal is to understand not just whether the AI answer appears, but whether it reinforces or weakens the brand narrative.

A practical measurement stack now needs at least four layers:

• Overview presence rate: How often does an AI Overview appear on your core branded queries?
• Brand mention and citation rate: When the overview appears, is your brand actually named and are your own pages cited?
• Source mix: Which domains and URLs keep showing up behind the answer?
• Message quality: Is the answer accurate, current, neutral, positive, or misleading?

That split matters because presence alone can be a false comfort. An AI Overview that appears on your brand search but leans on stale third-party pages is not a visibility win. It is a control problem. Likewise, a generated answer that mentions you but sends no click may still influence demand upstream of analytics.

This is where a dedicated measurement layer becomes useful. BotRank's AI Visibility feature helps teams run reusable prompts across multiple models, compare visibility trends over time, and see how brand mentions change by engine. If you want the reasoning behind that model, our piece on the AI visibility metrics most brands still miss explains why mentions, citations, rankings, and conversions should not be collapsed into one number.

The branded-query spike also reinforces a reporting truth many teams are still resisting: clicks are now a lagging indicator. If Google is answering first, a drop in branded CTR may show up after the real change has already happened. By then, the summary, the sources, and the user habit may already be established.

BotRank's Take: presence is not the same as control

The most important detail in the September data is easy to miss. DemandSphere counted an AI Overview whenever Google showed one, even if the brand itself was not cited inside the answer. That means a brand can “trigger” the AI layer and still lose the narrative to third-party pages, old help docs, or generic category content. For GEO teams, that is the real problem.

This is exactly where BotRank's Source Analysis becomes useful. It identifies the pages and domains that keep appearing behind AI answers, then lets teams verify whether those cited pages actually mention the brand and how useful those citations really are. That distinction matters. If your own site is shaping the answer, the fix may be a content or technical cleanup. If the answer is built from outside sources, the work shifts toward reputation, evidence, and broader entity clarity. A high presence rate looks impressive on a chart. Source control is what makes it operational.

What should brands do in the next 30 days?

They should treat branded AI Overviews like a live search surface, not a curiosity. The right response is not panic. It is a structured audit that turns a headline number into a list of concrete risks and opportunities.

• Build a branded query set. Start with your pure brand name, flagship product names, and the branded modifiers that matter most to revenue or trust. BotRank's Prompts Studio is useful here because it lets teams create reusable prompt sets instead of checking the same searches manually over and over.
• Capture the exact answer wording. Do not track only whether an overview appears. Save the text, note the placement, and compare week to week. A stable appearance rate can still hide a change in framing.
• Review the source stack behind the answer. If Google keeps leaning on old documentation, weak category pages, or third-party summaries, fix the inputs you control first. Then decide which external sources deserve active reputation work.
• Check tone and positioning. A generated answer can be factually correct and still frame the brand poorly. BotRank's Perception & Sentiment view helps teams analyze how the brand and its competitors are being described, which is often the missing layer in brand query reporting.
• Separate SEO reporting from AI answer reporting. Keep tracking rankings and branded clicks, but add a dedicated view for generated-answer presence, cited sources, and message quality. If you want the broader editorial playbook, our guide on how to earn AI mentions with GEO, not just rankings is a solid place to start.

There is also an organizational point here. Branded AI Overviews are not owned by one team. SEO may notice them first, but the fix can sit with content, product marketing, support, legal, documentation, or PR. If the generated answer summarizes your return policy, your product positioning, or your trust signals badly, the homepage alone will not solve it.

That is why this shift should be handled like a cross-functional visibility issue. The search interface changed, but the inputs behind it still come from many parts of the business.

FAQ: what does this mean for branded search?

Does an AI Overview on a branded query mean my brand is cited?

No. The September dataset counted an overview whenever one appeared, regardless of whether the brand itself was cited inside the answer.

Should I worry if the AI Overview appears lower on the page?

Yes, but in a measured way. Many observed branded overviews were mid-page rather than above the homepage, yet they still shape perception and can redirect attention to the generated layer.

Has Google explained the late-September jump?

Not publicly. The rise from 26.12% on September 1, 2026 to over 80% by September 28 happened without a corresponding Google announcement explaining the change.

What is the first thing I should audit?

Audit your top brand-name queries, the exact wording of the overview, and the sources behind it. That gives you a much clearer picture than rankings alone.

Is this only a Google issue?

No. Google is the immediate story here, but the broader lesson applies across AI search interfaces. Brand visibility now needs to be measured wherever generated answers are shaping the buying journey.

The old rule for branded search was simple: rank first for your own name. The new rule is harder and more important: make sure AI systems use the right evidence when they describe you. If Google is now writing the first line of your brand story, you need a way to monitor that story before the traffic report tells you something already changed.