On September 30, 2026, SEO consultant Chris Long reported that 93 of 100 enterprise brands he checked now trigger a Google AI Overview when you search the brand’s name, and Search Engine Roundtable’s Barry Schwartz confirmed the pattern the same day with screenshots for Reddit, Salesforce, Amazon, Adobe, Samsung, Apple and others. Rank-tracking vendor DemandSphere separately reported that AI Overviews on branded queries tripled during September.
That is a real shift. Earlier research from Ahrefs, which Long himself circulated, found AI Overviews skewed toward informational, non-branded, longer queries; a company’s own name was the one search most businesses assumed Google would leave alone. This piece reads the 93% figure carefully, explains what it does and does not tell a Michigan company, and ends with a test you can run before deciding anything about your branded Google Ads budget.
What did the September 2026 brand-query study actually find?
Short answer: AI Overviews appeared on 93 of roughly 100 large-brand name searches, but most sat below the official homepage rather than at the top of the page. Long collected the results with the SerpApi scraping service on a single morning. Schwartz noted that most of the AI Overviews showed “below the main company page or even lower down,” with Adobe a visible exception where the summary sat at the top.
Two exclusions also showed up: Google’s own name and news publishers did not get brand AI Overviews. Search Engine Roundtable itself did, which Schwartz attributed to the site having a Google Search Profile. Google has not published any statement about the change, so treat the scope and permanence as unconfirmed.
Does 93% mean your branded traffic is about to drop?
Not necessarily, and the study was not designed to say. It measured presence, not clicks. A count of “AI Overview on the page” tells you nothing about how many searchers still clicked the homepage sitting above it.
Three limits matter before anyone repeats the number in a budget meeting:
- The sample is enterprise brands. Reddit and Salesforce have millions of third-party pages describing them. A Livonia plastics molder or a Grand Rapids dental group has a few dozen. Whether Google triggers summaries as often for thin brands is unknown.
- One morning, one scraper. A SerpApi pull is typically a logged-out, single-location view. Logged-in mobile results in Michigan may differ.
- “Tripled” needs a base. DemandSphere’s figure is a growth rate; without the starting share it cannot be turned into an expected impact on your account.
My read: the headline overstates the click threat on the bare brand name and understates the narrative threat. If your homepage still ranks first and the summary sits underneath, most navigational searchers will click through as they always have. What changes is the story everyone else on that page reads about you.
How does Google build an AI Overview for a company name?
It summarizes what other pages say about the entity, then places that summary relative to how confident it is that the searcher just wants your website. That second part is the mechanism worth understanding.
For a bare brand query, Google’s strongest signal is navigational: the searcher wants your site. When that confidence is high, your homepage holds the top slot and the AI Overview gets pushed down, which matches the placement Long and Schwartz observed. The summary itself is assembled from cited sources, often directories, review sites, Reddit threads, job boards and news, not just your own pages.
Follow that logic to a case the study did not cover. If Google is less sure which entity you mean, because your name is generic, shared with another company, or recently changed, the navigational signal weakens. Expect the AI Overview to move up, and expect it to blend sources from more than one company. Adobe’s top-of-page summary hints that placement varies by brand even among household names.
Why are Michigan companies more exposed than the headline suggests?
Because a large share of Michigan business names are regionally generic, and the state’s supplier economy runs on buyers vetting vendors by name. Count the companies called Great Lakes, Wolverine, Huron, Superior, Lakeshore or Mitten something. Many have namesakes in Ohio, Wisconsin or Ontario. That is exactly the low-confidence entity situation where a summary can drift.
Now add the buying cycle. A purchasing manager qualifying a Tier 2 supplier in Sterling Heights or a stamping shop in Warren searches the company name, then “name + reviews,” “name + certifications,” and “name + careers” to judge stability. Consider a hypothetical: a Rochester Hills tooling firm shares its name with an Ohio shop that closed years ago. If the AI Overview blends the two and states the company “ceased operations,” the buyer never reaches the homepage to learn otherwise. No click is lost in the report; a quote request simply never arrives.
Multi-location service brands carry a related risk we covered when Google Business Profile panels started showing AI summaries: the local panel and the web-results summary can now describe the same business from different sources, and they may not agree.
Should you keep paying for branded Google Ads?
Keep them if competitors bid on your name or your homepage is no longer the first organic result; otherwise, test before you add budget. The brand AI Overview pushes third-party results and sitelinks further down, which makes the paid slot at the top relatively more visible. That is an argument for brand ads, but not proof they earn incremental clicks.
Michigan offers a clean way to test. The state splits into distinct Nielsen markets, including Detroit, Grand Rapids-Kalamazoo-Battle Creek, Lansing, Flint-Saginaw-Bay City and Traverse City-Cadillac, so a business serving more than one can pause brand ads in a smaller market and compare total branded clicks against a market where ads stay on.
How should a small team respond this month?
Audit what the summary says, fix the sources it cites, and measure branded clicks separately from everything else. All seven steps fit in a few hours a week.
- Build a brand query list. In Search Console’s Performance report, filter queries with a custom regex for your name and common misspellings. Keep the 10–15 variants with real volume, including “reviews,” “careers,” “pricing,” “near me,” and your name plus a city.
- Check each query in a private window on desktop and mobile. Record whether an AI Overview appears, whether it sits above or below your homepage, and every source it cites. Save screenshots with dates.
- Set a branded click baseline. Compare August 2026 with October 2026 for the brand regex only. Search Console does not break out AI Overview impressions, so the branded-only CTR trend is your best proxy. Our guide on diagnosing a traffic drop that isn’t a ranking drop walks through the comparison.
- Fix the cited sources first. If the summary pulls from a stale directory listing, an old Indeed profile or a Thomasnet page, update those before touching your own site. They are the inputs.
- Disambiguate your entity. Open your About page with one plain sentence stating legal name, city, founding year and what you make or do. Add Organization schema with
legalName,address,foundingDateandsameAslinks to your LinkedIn and directory profiles. - State the facts buyers check. Certifications such as IATF 16949 or ISO 9001, service areas, and locations belong in crawlable text, not only in a PDF or logo strip.
- Run the brand-ads split. Pause brand campaigns in one smaller Michigan market for 14 days, keep them on in Detroit, and compare combined paid plus organic branded clicks. Report errors in a summary through its feedback control, and keep a dated record.
What dates should Michigan marketers plan around?
- Early October 2026: Google’s September 2026 spam update began September 24 and was still producing ranking changes over the September 27–28 weekend, per Search Engine Roundtable. Hold any branded baseline until it settles so you are not measuring two changes at once. Our spam update guide covers how to separate them.
- October 13–27, 2026: A practical window for the 14-day brand-ads split, ending before holiday traffic distorts comparisons.
- November 26–30, 2026: Thanksgiving, Black Friday and Cyber Monday. Branded searches for retailers peak, so decide on brand-ad budgets before then.
- Early January 2027: Re-run the screenshot audit as B2B buyers return with new-year budgets and start vetting suppliers by name again.
Frequently asked questions about AI Overviews on brand searches
Can I stop Google from showing an AI Overview for my company name?
Not directly. There is no setting that removes AI Overviews from searches for your brand. Snippet controls such as nosnippet can limit how your own pages are quoted, but the summary can still draw on third-party sources, and restricting snippets can hurt your regular listing. The practical lever is improving the sources Google cites about you.
Will smaller Michigan businesses see brand AI Overviews too?
Possibly, but the September 30 test covered about 100 enterprise brands, so there is no published data for small and mid-sized companies yet. Check your own name in a private window on desktop and mobile. If a summary appears, note its position and sources; if none appears, re-check monthly, since DemandSphere’s data suggests coverage is still expanding.
How do I tell if the AI Overview is costing me branded clicks?
Filter Search Console to brand-name queries only and compare click-through rate for August and October 2026, after the spam update settles. If impressions hold steady while CTR falls and your average position is unchanged, something on the results page is absorbing clicks. Pair that with dated screenshots showing where the summary appears.
Does Google Search Profile affect brand AI Overviews?
Unconfirmed. Barry Schwartz said Search Engine Roundtable appeared to receive a brand AI Overview because it has a Google Search Profile, while other news publishers did not. Google has not explained the criteria. It is one data point worth watching, not a reason to create a profile purely to influence summaries.

Comments are closed