AI Overviews on branded queries tripled in September, tracking firm DemandSphere reported this week, as Google continued expanding the AI-generated answer surface to large brand names and even sports team rosters.

The DemandSphere read, summarized in the October 1, 2026 Search Engine Roundtable daily recap, marks one of the sharpest monthly shifts for AI Overviews since the feature's general rollout. Google tests now surface AI Overviews for "almost every large brand name except for Google's own name and news publications," according to the recap.

What does the team-roster change mean for publishers?

The brand-name expansion follows a separate test that pulls AI-generated answers into team-roster queries. Searchers looking up specific rosters receive an AI Overview that aggregates content typically sourced from team websites, league databases and Wikipedia, displacing those links further down the page.

The change adds a sports-reference vertical to a list that already includes health queries and product research. Publishers running roster templates should track click-through rates and AI Overview presence over the next several weeks to gauge the vertical impact.

What changed at Microsoft Advertising

Microsoft moved optimization experiments to general availability on October 1. The tool runs controlled A/B tests that measure the impact of one campaign change inside Search, Shopping, Audience and Performance Max campaigns, Microsoft's help center explains. Advertisers in those four campaign types can now set up an experiment without an enrollment step.

What shipped in ChatGPT Ads

OpenAI emailed advertisers this week with a package of ChatGPT Ads updates:

• Bulk creation for product feed campaigns
• A product review status field
• API reporting that captures conversions beyond a single campaign goal
• Separate account, brand and legal name fields
• Live inventory in Southeast Asia and Taiwan

The Southeast Asia and Taiwan launch documents the first regional expansion of ChatGPT Ads beyond its prior markets, per the recap.

Is the Local Service Ads bug confirmed?

Google Local Service Ads is serving ads from advertisers who do not target the queried location, according to multiple reports in the Search Engine Roundtable thread. The behavior appears to override geo-targeting settings advertisers set inside their accounts. Impact concentrates on home-services advertisers in metropolitan markets where LSA spend runs heaviest.

Google has not commented publicly on the bug. Advertisers running LSA should compare served locations against their configured target list and flag mismatches to Google Ads support.

What did Google change in the destination policy?

Google quietly added a line to its destination-not-working policy document that now reads: "Make sure your campaign has at least one asset group configured. If an asset group is missing, you may receive an 'invalid final URL' error during setup even if your destination URL is working correctly."

Google said the change "provides additional examples for destination issues for advertisers." Advertisers rebuilding Demand Gen or Performance Max campaigns should confirm at least one asset group exists before launch.

Which UI tests is Google running?

Google is running two visual experiments on the main search bar. In one, the "I'm Feeling Lucky" button disappears once a searcher begins typing. In a second variant, the button is supplemented with AI-mode shortcuts. The test echoes an earlier experiment Google ran and then walked back.

A separate Shopping test displays three products in the "more products" grid and can expand to four or five items depending on viewport width. Merchants running free listings should monitor whether the new grid size changes product discovery and CTR.

Related moves around the industry

The September 2026 Google spam update "started off strong and is still rolling out," per the October 2026 webmaster report, putting low-quality AI content sites and templated directory pages in its path. Lily Ray wrote on X that affected sites "are actually producing many of the exact same page types and templates that companies scaled with freelance writers before 2023."

Glenn Gabe, also on X, noted roughly 6 million URLs indexed inside one Facebook Answers capture, all of which "either the spam update hammered the directory or Google was displaying altered SERPs."

Separately, OpenAI's chief research officer told MIT Technology Review the company will not "shoot ourselves in the foot" in responding to recent hack fallout. Google introduced Gemini 4 this week, with CNBC reporting a multi-week company-wide testing period preceded the launch.

What to monitor next

Watch the September 2026 spam update's completion window and any confirmation from Google about whether the AI Overview brand-name expansion is a global rollout or a regional test. DemandSphere's next monthly read, due in October, will show whether the September tripling on branded queries holds.