AI Overviews Now Appear in 83% of Branded Searches, Ahrefs Finds

The September 2026 spam update — Google's fourth spam update of the year, first reported in the September 21st issue as a global two-week rollout with early volatility on September 23rd–24th — has now delivered two distinct impact waves. The first wave struck over the weekend of September 25–27, with significant movement observed on Saturday and Sunday. After a brief lull, a second wave hit on September 30, prompting renewed turbulence across the SEO community. Search Engine Roundtable's weekly video recap confirmed both windows and noted the update as the dominant story of the week. The rollout has not yet concluded, and further changes remain possible.

Google made substantial updates to its 'Creating helpful, reliable, people-first content' documentation this week, adding two significant new sections. The first emphasizes that a page's main content is critically important, defines what constitutes main content, and explains how Search Quality Raters are trained to evaluate it. The second addition is an explicit warning against fake or fabricated author profiles, a signal that author authenticity is now a formal quality concern in Google's official guidance — not just in rater guidelines.

Marie Haynes called this her favorite piece of Google search documentation and noted it directly reflects the principles behind Google's ML ranking systems. She speculated that the update could signal an imminent core algorithm update, potentially tied to advances in Google's Gemini models. Search Engine Journal also flagged the fake-author warning as a prompt for site owners to audit author bios and credentials. Google additionally updated its AI-generated content guidance in parallel, now explicitly requiring publishers to manually fact-check all AI content before publishing — including titles, alt text, and metadata, not just body copy.

Ahrefs published original research this week showing that AI Overviews now appear for 83% of branded search queries, diverting clicks away from brand-owned websites toward third-party sources such as Wikipedia and YouTube. Search Engine Roundtable and Search Engine Land corroborated the trend, with SER observing that Google is now showing AI Overviews for almost every large brand name — with the notable exceptions of Google itself and news publications.

The implication for brand SEO is significant: even users actively searching for a company by name may now receive an AI-generated summary before ever reaching the brand's own website. Google is also expanding AI Overviews into structured, data-heavy content like team sports rosters, generating its own AI answer rather than linking to the original source — a direct threat to niche content publishers in those verticals.

Google is testing URL tracking parameters on links within AI Overviews and AI Mode, which would allow site owners to distinguish clicks coming from AI Overviews from standard organic clicks in analytics — a long-requested capability. Search Engine Roundtable reported the test and noted that if fully rolled out, it would meaningfully improve traffic attribution for publishers currently blind to their AI Overview performance.

Separately, Google Gemini has already begun appending UTM parameters to outbound links, confirmed by Search Engine Journal and SEJ's SEO Pulse roundup. This gives site owners cleaner referral attribution for traffic originating from Google's AI platform. On the UI side, Google is once again testing citation cards displayed at the bottom of AI Overview responses on desktop rather than on the right side — a layout variation trialed at least twice before, which directly affects click-through visibility for cited sources.

The AI Contribution Pilot — first spotted in April 2026, confirmed with detail in the September 14th issue as a small test compensating publishers when their content appears in AI Overviews, AI Mode, and Gemini — has now revealed its actual payout scale. According to a report cited by both Search Engine Roundtable and Search Engine Journal, Google is paying roughly 100 publishers approximately one-tenth of 1% of their advertising revenue. SEJ added that payouts vary widely between participants and some publishers still cannot determine how their amounts are calculated, raising serious questions about program transparency and sustainability.

Separately, a Penske Media lawsuit (Rolling Stone, Variety, Billboard, The Hollywood Reporter) against Google over AI Overviews was dismissed this week, with the judge finding no formal agreement existed between the parties — closing off one potential legal avenue for publishers seeking compensation.

Several pieces this week addressed how to earn and influence citations in AI-powered search results. An AnswerShare case study reported in Search Engine Journal showed that publishing structured, sourced brand context measurably increased brand recommendations in AI search outputs — an early data point for the emerging GEO (Generative Engine Optimization) discipline. Ahrefs published a practical GEO workflow covering how to diagnose missing brand mentions, wrong topic associations, and access barriers for AI crawlers.

Lumar's ongoing GEO series argued that ranking well in traditional search does not guarantee eligibility as an AI source, identifying six practical checks — including crawlability, content structure, and entity clarity — that determine whether a page is even considered for AI-generated answers. Meanwhile, an SEJ column argued that Wikipedia, podcasts, and Reddit are now the primary drivers of AI citations, and that SEOs need to collaborate with PR teams to build these third-party signals at scale. A separate analysis from Alek Asaduryan noted that a Neil Patel bracket survey of 512 marketers crowned Digital PR as the top single GEO tactic, though the author emphasized that stacking multiple tactics is necessary to earn consistent citations from ChatGPT, Perplexity, and Google AI Overviews.

A significant piece of evidence entered the public record this week: a court filing quoted an OpenAI engineer stating that users 'won't click' links in AI chat interfaces, supported by Microsoft data showing lower CTRs in Bing Chat than in traditional search. Reported by Search Engine Journal, this is one of the clearest first-hand admissions that AI search inherently suppresses click-through to external websites — a trend SEOs have been tracking empirically but rarely seen confirmed in legal filings.

Complementing this, Q2 2026 data analyzed by Search Engine Journal showed that Google desktop CTR declined while mobile CTR rose at top search positions. The analysis spans the period covering the end of a known Search Console logging error, adding a measurement complexity caveat. Growth Memo's Kevin Indig (co-developed with a VP of Growth at Ramp) published a broader argument that traditional attribution models are fundamentally broken for the AI era — clicks and cookies were always a poor proxy for non-paid marketing value, and with both fading, the industry needs new budget allocation frameworks. He suggests AI visibility is now a core channel that requires its own measurement logic rather than being forced into cookie-based attribution.

As AI Overviews become a dominant SERP feature, the industry is grappling with how to turn raw visibility data into actionable business metrics. Search Engine Journal ran a practical critique of Google's Generative AI report in Search Console, arguing it shows only visibility data with no commercial context — the fix proposed is layering in traffic dependency and commercial value metrics to create real exposure analysis.

Cyrus Shepard joined SearchPilot to discuss a critical nuance: winning AI citations does not equate to winning in traditional search, with the two surfaces now requiring distinct strategies. A separate SEJ piece outlined five metrics practitioners are using to track AI search's contribution to client pipelines: brand presence, answer accuracy, citation share, and attribution. Lumar's eligibility framework adds that crawlability and structured content are near-binary prerequisites for even being considered as an AI source.

Google globally rolled out its AI Mode information monitoring feature to all users this week, having previously limited it to paid Ultra and Pro subscribers. The feature lets any Google user set up ongoing monitoring tasks and receive updates within Search — effectively turning one-time queries into persistent background agents. Both Search Engine Journal and Search Engine Roundtable confirmed the expansion.

Search Engine Journal published a wider analysis of this behavioral shift: as Google AI Mode and OpenAI's 'Dots' feature both enable queries to become 'standing instructions' that run continuously, a fundamental change in how discoverability works is underway. Unlike transactional queries that end with a click, persistent agents repeatedly query sources over time — raising new questions about site visibility, server load, and traffic attribution that the industry has not yet fully worked through.

Google announced it will adopt the Universal Content Protocol (UCP) draft specification for hotel booking within AI Mode over the coming months, signaling a move toward open standards for structured commerce in AI search — a notable development for travel and hospitality SEOs and structured data practitioners.

Google also expanded its Loyalty Customer Match program into AI Mode and Gemini, as well as additional geographic regions, per updates to Google Merchant Center documentation. Google Ads sitelink formats are also being tested within AI Mode as Google works to determine which ad formats generate the best CTR in the new AI-powered surface — an early indication of how Google plans to monetize the channel. Combined with Google's test of dropping the 'I'm Feeling Lucky' button in favor of AI-oriented alternatives, the signals this week point to an accelerating transition from traditional search to AI-first interfaces.

Following the September 7th issue's report of Google's extraordinary public statement to Reuters — that DMA compliance represents "the largest reduction in quality of service" in its 29-year history — the story has now escalated further. This week, Google SVP of Knowledge & Information Nick Fox made on-record comments to interviewer Pietro Valetto characterizing the DMA-mandated changes as the biggest quality reduction in the company's history and saying implementation made him feel 'terrible' and 'sad' — adding a senior executive's personal voice to what was previously a corporate statement.

Separately, Google has formally appealed the European Commission's requirement to share search history data with competitors, citing inadequate anonymization safeguards and risks to Android security protections. The appeal covers two distinct EC decisions and carries broad implications for Google's data ecosystem and competitive dynamics in EU search markets.

Google's updated AI-generated content guidance — which now explicitly requires manual fact-checking of all AI content before publishing, extending to titles, alt text, structured data, and metadata beyond body copy — was flagged this week by both Search Engine Journal and Search Engine Roundtable. The policy update arrives alongside new data from an Ahrefs survey of 301 marketers showing that 93% now use AI in content creation, with 69% publishing more AI content than a year ago and Claude emerging as the most-used model. The combined picture: AI content production is now effectively ubiquitous, and Google is raising its quality floor in direct response. Note that Google's AI content guidance update is also covered in the 'Helpful Content Guidelines' section of this issue in the context of its broader documentation revision.

Cloudflare moved its Monetization Gateway and Pay Per Use products into closed beta, introducing a mechanism where AI agents are charged per request and publishers can earn revenue when buyers report AI usage of their content. This is one of the most concrete infrastructure-level moves yet to establish an economic model for AI-driven content consumption.

The development arrives in the context of a broader publisher squeeze: Amazon blocked Meta's AI shopping agent Muse using server-side detection rather than robots.txt, highlighting a major gap in traditional crawl-control mechanisms as AI agents become harder to detect and block. Search Engine Journal noted that the fundamental challenge is not writing a robots.txt rule — it's identifying agents that look like human shoppers. Together, these stories trace the emerging battle lines between content owners, AI platforms, and the infrastructure layer trying to mediate between them.

A series of notable changes and bugs hit local search this week. Google Maps now requires users to sign in to a Google account to view all reviews on a Business Profile, sort reviews, or leave a review — a change Search Engine Roundtable attributes to efforts to prevent AI and bot scraping. Local SEO monitoring workflows that rely on unauthenticated review access will need to be updated.

A separate Google Maps bug is incorrectly placing map pins for some service-area businesses in the middle of the ocean, affecting any GBP listing that does not publicly display a physical address. Local SEOs should audit client listings and submit reports through Google's support channels. Additionally, a bug in Google Local Service Ads was found serving ads to geographic locations advertisers have not targeted — potentially wasting spend and degrading user relevance. Google also updated its Google Business Profiles policies to add a new 'Misuse of support channels' section.

Google is testing a new sponsored products grid format that visually mimics the organic product listing grid, replacing the traditional horizontal carousel — a change that could significantly blur paid/organic distinctions for users and affect e-commerce CTR dynamics.

On the AI Overviews UI front, Google is experimenting with replacing the 'Show more' expand button with a 'Loading…' animation, and has previously abandoned a test of 'Enter AI Mode' as the label. These UI experiments add to the citation card layout variation (bottom-of-response on desktop) reported separately in the AI Overviews section of this issue. Sitelinks are also being tested in new line-by-line mobile layouts, and Google is testing an expandable product grid in regular results that shows three products initially before expanding. Taken together, these experiments reflect simultaneous optimization of SERP layouts for paid performance and AI integration.

Building on the September 21st issue's report that Google auto-enables native checkout in AI Mode via UCP (with opt-out required), agentic commerce infrastructure expanded significantly this week on two fronts. Shopify extended its WebMCP tools into the checkout process, enabling browser-based AI agents to update carts and place orders after buyer confirmation. Shopify also introduced an 'agentic storefronts' setting that auto-enrolls merchant stores into new AI shopping channels by default — including Meta and unnamed startups — meaning merchants must actively opt out and audit which platforms receive their product data.

On Google's side, its Loyalty Customer Match program expansion into AI Mode and Gemini and the announced adoption of the UCP hotel booking specification (covered in the AI Mode section of this issue) further embed agentic commerce into AI search infrastructure. Steve Toth's SEO Notebook flagged Google's emerging standard for enabling AI agents to transact directly on websites — a development with direct implications for how product pages are structured and optimized.

Google Analytics experienced a confirmed outage this week, with the platform failing to load for users before Google's status page indicated a staged restoration — 'restored for some users' with a full fix 'expected in the near future.' The incident serves as a reminder of SEO teams' dependency on GA for daily reporting.

In a separate update, Google Analytics now supports app conversions in its cross-channel performance and attribution reports, expanding measurement capabilities for marketers managing campaigns across web and app environments. For teams running multi-channel campaigns, this closes a longstanding gap in cross-channel attribution within GA4.

Publishers had a difficult week in court. A federal judge dismissed antitrust claims brought by publishers against Google, acknowledging that real harm to publishers exists but ruling their legal arguments did not meet the antitrust threshold. The Penske Media lawsuit over AI Overviews (Rolling Stone, Variety, Billboard, The Hollywood Reporter) was also dismissed — the judge found Google had made no formal enforceable agreement with Penske. (The Penske dismissal is also detailed in the AI Overviews section of this issue in the context of publisher compensation.)

On the WordPress front — following the September 7th issue's report that Automattic's board placed Matt Mullenweg on paid leave amid conspiracy claims — a judge this week refused to dismiss WP Engine's antitrust claims against Automattic and Mullenweg, while Automattic also secured seven partial wins in the same ruling. The dispute continues to have implications for the WordPress plugin and theme ecosystem that a large share of SEO infrastructure depends on.