The Search Signal

AI got more selective this quarter. Victorious went back to the same 161 brands from its Q2 Quarterly Search Report, asked nine AI platforms the same questions, and found fewer brands named and a shorter list of sources behind the answers. The brands that kept their place had one thing in common: AI was already citing their content while buyers were still researching the category, before anyone asked who to hire. Michael walks through the Q3 findings, what changed since July, and how to plan around one quarter of data.In this episode:Brands AI cited on research questions were named on consideration questions 94% of the time, versus 50% for brands it didn’t citeAI named 29 brands this quarter, down from 42, with most losses among brands with 2,000 to 10,000 web mentionsThe 50-50 threshold for getting named rose from about 20,000 mentions to 24,000 to 28,000Brand websites lost citation share on research questions to government and industry association sitesChatGPT cited Reddit in 220 of 1,000 answers last quarter and zero this quarterAI Overviews appeared on 97.5% of tracked Google searches, up from 73%, and 37% of their sources rank in the top 10Why three new AI traffic studies measure the wrong thingFind the full transcript and links to sources on the episode page: https://victorious.com/podcast/ai-recommends-cited-brands/Links:Victorious: https://victorious.comMichael Transon on LinkedIn: https://www.linkedin.com/in/michaeltranson/Victorious on LinkedIn: https://www.linkedin.com/company/victorious-seo/Chapters:00:00 Why AI recommends the brands it already cites01:33 Google ranking volatility: unconfirmed update or bad rank tracker data?02:51 PHP security flaws and patching WordPress sites03:28 Three AI search traffic studies and the measurement problem they ignore06:26 Q3 2026 Quarterly Search Report: what changed since Q208:36 Methodology: 161 brands, nine AI platforms, research vs. consideration questions11:05 Headline finding: AI cites brands at the research stage before recommending them14:43 Government and industry association sites take citation share from brand websites18:25 AI named fewer brands and the mention threshold went up20:33 What predicts AI brand mentions: web footprint, not link authority23:23 AI citation sources consolidate in legal, healthcare, and financial services24:59 ChatGPT drops Reddit citations and AI platforms disagree on sources28:11 Google AI Overviews reach 97.5% of searches and still follow rankings31:36 Should you change your search strategy on one quarter of data?35:32 What Victorious is tracking for the Q4 Quarterly Search Report37:09 Key takeaways from the Q3 report

Google’s Universal Cart went live in the US on October 15, putting one shopping cart across Search, AI Mode, and the Gemini app. Michael spent time with it before recording, saving products from different stores and following them to checkout to see where it works and where it breaks. He covers whether this counts as agentic shopping, why Google wants the experience to be seamless, what Google reads when it compares your products, and how to get set up before Black Friday.In this episode:What the cart does, from price-drop alerts to the two ways it handles checkoutWhy Google puts the cart on its shopper “confidence and control” track, and what ChatGPT’s checkout results say about trustWho it matters for, including Shopify stores that were enrolled by defaultWhat Google sees in the cart that your analytics no longer doesThe product data Google reads, including Merchant Center’s new conversational attributesA Black Friday setup plan and a 10-minute test you can run todaySix predictions for where the cart goes nextChapters:[00:44] Intro[01:29] News: Google serves altered results to rank trackers[03:39] News: ChatGPT visits landing as direct traffic[05:31] News: Google tests ads in AI Mode for connected apps[07:06] News: The AI Mode sidebar shows up on regular search[07:47] Google’s Universal Cart goes live[08:56] What launched[12:13] Is this agentic shopping?[17:23] Who Universal Cart matters for[20:35] Why Google wants the cart to be seamless[23:11] What Google reads when it compares products[26:37] How to set up before Black Friday[29:58] Test it the way a shopper would[30:49] Six predictions[33:50] The takeawayThe transcript and links to all resources can be found on the episode page: https://victorious.com/podcast/universal-cart/Sources:Google, An intelligent new shopping cart (Universal Cart launch), October 15, 2026: https://blog.google/products-and-platforms/products/shopping/universal-cart-launch/Retail Brew, Google set to launch a Universal Cart (I/O 2026 preview): https://www.retailbrew.com/stories/google-takes-agentic-shopping-to-youtube-set-to-launch-a-universal-cartPPC Land, Google feeds rank trackers swapped results on 8% of pages, Nozzle finds: https://ppc.land/google-feeds-rank-trackers-swapped-results-on-8-of-pages-nozzle-finds/Search Engine Roundtable, OpenAI drops UTM parameters from ChatGPT results (GPT-6): https://seroundtable.com/utm-parameters-missing-chatgpt-42299.htmlSearch Engine Roundtable, Google ads now in AI Mode with Personal Intelligence: https://seroundtable.com/google-ads-ai-mode-personal-intelligence-42320.htmlGoogle tests AI Mode left sidebar on main search results: https://seroundtable.com/google-ai-mode-left-sidebar-42298.htmlNick Fox on what AI search agents and Universal Cart mean for marketers: https://ppc.land/nick-fox-on-what-ai-search-agents-and-universal-cart-mean-for-marketers/Search Engine Land, Walmart: ChatGPT checkout converted worse than its website: https://searchengineland.com/walmart-chatgpt-checkout-converted-worse-472071What went wrong with ChatGPT's Instant Checkout: https://www.modernretail.co/technology/what-went-wrong-with-chatgpts-instant-checkout/Inside the Worldpay agentic commerce report: https://www.worldpay.com/insights/articles/inside-worldpay-agentic-commerce-reportSearch Engine Journal, Google's Universal Cart is live: how carts get to your store: https://www.searchenginejournal.com/google-universal-cart-store-handoff/593145/PPC Land, 8 Google Merchant Center attributes your feed needs for AI Mode: https://ppc.land/8-google-merchant-center-attributes-your-feed-needs-for-ai-mode/

Anthropic updated its usage policy on October 8 to ban using Claude to build networks of fake sources that sway search results and AI answers. The rule takes effect November 12 and reaches every product built on Claude, which means a vendor running that playbook on your behalf puts your whole account at risk. Michael walks through why he reads this as good news, how the case that prompted it got caught, where faking still works, and what to build instead.In this episode:What the new usage policy section on deceptive campaigns prohibits, and who it applies toThe 70-site fake local news network that got caught by its hosting, registration dates, and publishing timingWhy more automation makes a network easier to spot, not harderHow you'd find out if a platform discounted sites that mention you (short answer: you probably wouldn't)Where faking still works, and why data voids are the real exposure for small and new companiesFour things to do this quarter, including the questions to ask any AI visibility vendorChapters:[00:45] Intro[01:57] Why Michael is happy about this[03:20] How AI answers get built, and why corroboration is the target[04:57] What changed in Anthropic's usage policy[09:46] The 70-site network and how it got caught[12:43] Platforms closing in from both sides[15:31] How you'd even know your brand was affected[18:15] Where faking still works: NewsGuard, Harvard, and Cornell Tech[22:06] What a durable presence looks like[24:27] Four things to do now[31:25] The November 12 deadline[33:06] OutroFind a complete list of linked sources at https://victorious.com/podcast/anthropic-bans-fake-sources/Sources:Anthropic, 2026 Usage Policy update: https://www.anthropic.com/news/2026-usage-policy-updateAnthropic, Usage Policy: https://www.anthropic.com/legal/aupAnthropic, Detecting and countering misuse of AI, September 2026: https://www.anthropic.com/threat-intelligence-report-september-2026OpenAI, Disrupting AI-enabled false front operations: https://openai.com/index/disrupting-ai-enabled-false-front-operations/Google Search spam policies: https://developers.google.com/search/docs/essentials/spam-policiesGoogle Search ranking updates: https://developers.google.com/search/updates/rankingGoogle, Using generative AI content: https://developers.google.com/search/docs/fundamentals/using-gen-ai-contentGoogle, Creating helpful, reliable, people-first content: https://developers.google.com/search/docs/fundamentals/creating-helpful-contentGoogle, AI features and your website: https://developers.google.com/search/docs/appearance/ai-featuresReddit bot labeling and human verification (TechCrunch): https://techcrunch.com/2026/03/25/reddit-bots-new-human-verification-requirements/NewsGuard, Pravda network audit: https://www.newsguardtech.com/special-reports/moscow-based-global-news-network-infected-western-artificial-intelligence-russian-propaganda/Alyukov et al., HKS Misinformation Review: https://misinforeview.hks.harvard.edu/wp-content/uploads/2025/10/alyukov_chatbots_kremlin_disinformation_20251015.pdfZhang, Triedman, Shmatikov, Cornell Tech (preprint): https://arxiv.org/pdf/2605.24245FTC final rule on fake reviews: https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonialsShared ChatGPT conversations in Search Console (GSQI): https://www.gsqi.com/marketing-blog/chatgpt-links-in-google-search-console/

AI assistants decide whether to name your company based mostly on what other websites say about you. This week Michael lays out the three programs Victorious runs to get more of those websites talking, and the two things that have to be in place first.In this episode:* The three programs, broadest to most targeted: mention building, link building, citation outreach* Why mentions beat links for AI visibility, per Ahrefs's 75,000-brand study* The two foundations: an in-house expertise program, and co-occurrence* Michael's case for paid placements, and the one word in Google's guidance it rests on* How citation outreach starts: 20 to 50 buyer prompts and a monthly list of 50 pagesFind all the sources referenced in this episode on the Victorious website: https://victorious.com/podcast/mention-building-program/Chapter Headings0:00 Intro1:32 News: Google says to review AI content before publishing3:45 Google is paying some publishers for AI answers4:56 AI Overviews on most branded searches6:32 The three programs, defined8:24 Why mentions beat links9:18 Foundation one: an expertise program13:07 Foundation two: co-occurrence15:44 Mention building: nine ways to get on other sites18:37 Press releases20:53 The case for paid placements23:25 Link building: the four-level review26:07 Citation outreach27:28 Four kinds of citations28:45 Working the top 50 pages every monthSourcesGoogle Search Central, "Guidance on using generative AI content," updated October 1, 2026, https://developers.google.com/search/docs/fundamentals/using-gen-ai-contentSearch Engine Roundtable, "Google: Manually Factcheck & Review AI-Generated Content For Accuracy," October 1, 2026, https://www.seroundtable.com/google-updates-ai-content-guidelines-factcheck-review-42217.htmlSearch Engine Roundtable, "Google Paying ~100 Publishers 0.1% Of Ad Revenue In AI Contribution Pilot" (citing The Information, September 29, 2026), September 29, 2026, https://www.seroundtable.com/google-ai-contribution-pilot-01-percent-42188.htmlDemandSphere, "AI Overviews on branded queries tripled in September," October 1, 2026, https://www.demandsphere.com/blog/branded-ai-overviews-september-2026/Ahrefs, "AI Overviews Currently Appear for 83% of Branded Searches," October 2, 2026, https://ahrefs.com/blog/ai-overviews-on-branded-searches/Ahrefs, "An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)," https://ahrefs.com/blog/ai-overview-brand-correlation/Victorious, Q2 2026 Quarterly Search Report, https://victorious.com/quarterly-search-report/BrandMentions, 2026 AI visibility study (240 brands, 410,000 mentions), https://brandmentions.com/blog/?p=4770Octolens, "How B2B SaaS Wins on Reddit: 522M Mention Analysis," 2026, https://octolens.com/reddit-b2b-saas-studyMuck Rack, "7 FAQs about the May 2026 update of 'What Is AI Reading?'," 2026, https://muckrack.com/blog/what-is-ai-reading-new-insights-may-faqsGoogle Search Central, "Optimizing your website for generative AI features in Google Search," May 2026, https://developers.google.com/search/docs/fundamentals/ai-optimization-guide

When marketing leaders ask whether their technical SEO covers them for AI search, Michael's answer is mostly yes. Google's AI Overviews and AI Mode run on Google's regular index, and most of a standard technical audit carries over to the other platforms. This week he sorts technical work into what to keep, what to start, and the short list of things to skip, including a training-bot decision that removes your own description of your company from what the next model learns.In this episode* Why indexability now means two things: a search index and a model's training data* How OpenAI and Anthropic split their crawlers by job, and why blocking the training bot still costs you in answers that don't involve a search* Which index each platform searches: Google's for AI Overviews and AI Mode, a mix led by OpenAI's own index for ChatGPT, Brave for Claude, and Bing for Copilot* Two Google details teams miss: Google-Extended doesn't affect AI Overviews, and old nosnippet tags can keep pages out of them* The six areas of a technical audit that carry over, and why schema facts also belong in visible text* Why every important page should make sense as the only page an assistant reads* How to block training narrowly by folder, and why robots.txt is a request your server may not enforce* Why Michael recommends skipping LLMs.txt for most sites* What could change: less live search, paid web search, agents that read pages visually, and agentic checkout* The order to work through it with your teamChapters[00:00:00] Intro[00:02:26] This week in search: Google's September 2026 spam update, Cloudflare's AI crawler defaults, Google's Web Search Service API docs disappear[00:05:33] Does your technical SEO cover AI search?[00:06:04] Keep, start, stop[00:07:33] Crawlability and the two kinds of indexability[00:09:31] Separate bots for training and search[00:10:32] AI Overviews and AI Mode: Google-Extended and snippet controls[00:12:11] ChatGPT's own search index[00:14:13] Claude, Brave, and Copilot[00:15:04] What to do for each platform[00:17:00] Keep: the six areas of a technical audit[00:17:54] Site architecture and internal linking[00:20:16] What your server logs can tell you[00:20:42] Rendering and page speed[00:22:00] Structured data[00:23:47] Start: make every page make sense on its own[00:25:42] Start: make crawler decisions on purpose[00:28:40] Blocking training narrowly[00:29:12] Robots.txt is a request[00:30:58] Stop: LLMs.txt[00:33:46] Every black hat tactic was once white hat[00:34:38] What could change[00:39:13] Where to start with your teamFind all the resources from this episode on our website.SourcesSearch Engine Journal, "Google Rolls Out September 2026 Spam Update Globally," Sept. 2026, searchenginejournal.com/google-september-2026-spam-update/590828/The Next Web, "Cloudflare gives AI crawlers a September deadline: pay publishers or get blocked," July 2, 2026, thenextweb.com/news/cloudflare-block-ai-crawlers-pay-publishersGoogle Search Central, "AI features and your website," developers.google.com/search/docs/appearance/ai-featuresOpenAI, "Overview of OpenAI Crawlers," platform.openai.com/docs/botsPeec AI, "ChatGPT built its own search index," Sept. 2026, peec.ai/blog/chatgpt-built-its-own-search-indexHasData, "AI Crawler Block Index," hasdata.com/blog/ai-crawler-block-indexForrester, "What It Means That The Leader In 'Agentic Commerce' Just Pulled Back," March 2026, forrester.com/blogs/what-it-means-that-the-leader-in-agentic-commerce-just-pulled-back