Google's Head of Search Doesn't Know Where Search Is Going

AD

Turn AI into Your Income Engine

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

• A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential

• Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background

• Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

SEO + GEO

Google's Liz Reid: Search and Gemini May Converge or Diverge Further

Liz Reid, VP and head of Search at Google, drew a clearer line between Search and Gemini but said the long-term direction is still open. In an interview on Access Podcast, she said Search is an information product focused on connecting people with the web, while Gemini centers on productivity and creation. She acknowledged the boundaries are fluid and the eventual outcome could be convergence, divergence, or a third product altogether as agents reshape internet interaction. She expects a future with more agent-to-agent communication rather than just humans browsing directly, and said the company wants to do more to surface Trusted and Subscription-based Sources more prominently, pointing to the Preferred Sources feature as a step toward prioritizing content users already pay for.

Keyword Research for GEO: The FAN Methodology

Limor Barenholtz introduces the FAN Methodology for GEO keyword research, suggesting that traditional keyword-to-page mapping structurally fails in LLM search because queries are decomposed into 6 to 20 sub-queries before any content is retrieved.

• AI search queries average 70 to 80 words compared to 3 to 4 words in traditional Google search

• 35% of consumers use AI tools at the discovery stage versus 13.6% using traditional search.

• Only 27% of fan-out sub-queries remain consistent across different searches of the same query, and 66% appear only once, making optimization for fixed sub-query lists a moving target.

• The FAN framework has three components: Fan-out Mapping (decompose anchor queries into sub-query space), Authority-Signal Alignment (add quantified data), and Node Architecture (structure content as independently retrievable chunks).

• SEOClarity analysis found 25% of ChatGPT's top 1,000 cited URLs had zero organic visibility on Google, and only 12% of URLs cited by ChatGPT, Perplexity and Copilot rank in Google's top 10.

• Every anchor query fans out across seven sub-query types: Definition, Comparison, How-to, Use case, Objection, Entity expansion, and Metric.

• Gemini, GPT and and Claude share 78 to 84% of content preference rules, meaning citation-worthiness signals are largely consistent across platforms.

My Take: I had a serious sense of déjà vu reading this. I kept thinking - where have I read this before? Oh right, I wrote about it two weeks ago in my ARENA Framework guide for AI Search.

It's not identical, but the core concepts overlap heavily, especially around what I call "retrieval assets" and building a portfolio of pages targeted at the sub-questions AI systems fan out into. SimilarWeb's piece even uses the same page-type taxonomy: definition, comparison, criteria, and objection pages. Same idea, different label.

To be clear, I'm not accusing anyone of anything. These ideas are in the air right now, and smart people are converging on the same conclusions independently. But it is validating to see a well-resourced team at SimilarWeb land in the same place a couple of weeks later. A couple others below also mention the same concepts in their posts.

The difference is mostly in form. Their article is a full methodology guide - data-heavy, researcher-cited, with a named framework (FAN), a measurement stack, and platform-by-platform breakdowns. Mine is a playbook: skip the justification, here's what to build and how to link it. Both are worth reading. But if you want to act on it today, the playbook is the faster on-ramp.

Their article is a full methodology guide - data-heavy, researcher-cited, with a named framework (FAN), a measurement stack, and platform-by-platform breakdowns. Mine is a playbook. Skip the justification, here's what to build and how to link it. If you want to act on it today, the playbook is the faster on-ramp. Both are worth reading.

Reddit's AI Citation Decline: How Brands Win Back Visibility

Conductor analyzed 238,212 prompts across LLMs and found that Reddit's overall AI citation share dropped 50% between October 2025 and January 2026, falling from 2.02% to 1.01%. But when Reddit is cited, it increasingly owns the conversation.

• Reddit's share of responses where it is the only cited source increased 31% from October 2025 to January 2026.

• Of 145,662 classified queries where Reddit was the only source, Transactional prompts led at 36.5%, followed by Commercial at 33.5%, Informational at 25.6%, and Navigational at just 4.4%.

• Reddit wins for lived experience and peer perspective because brand content rarely shows up with better answers.

• Reddit dominates where users want unfiltered opinions, visible disagreement, and failure stories.

• Brand content wins on authoritative medical answers, financial explainers, reference facts, and structured how-to tutorials.

38% of AI Overview Citations Pull From Top 10 Pages

Louise Linehan analyzed 863,000 keyword SERPs and 4 million AI Overview URLs to update Ahrefs' prior research showing that 38% of pages cited in AI Overviews also rank in the top 10 organic results, down from roughly 76% in July 2025.

• The remaining citations split almost evenly between positions 11 to 100 (31.2%) and pages beyond the top 100 (31.2%).

• YouTube now accounts for 18% of non-ranking citations, making it the most cited domain in AI Overviews and growing 34% over the last six months.

• AI Overviews are relying more on query fan-out

• Ranking for the original query is no longer enough to win citations. Google is pulling from fan-out query results where your page may not rank at all.

Google Patent: Sending Searchers to AI-Generated Pages Over Your Site

Barry Schwartz covers a Google patent named "AI-generated content page tailored to a specific user" that describes generating custom landing pages and sending searchers to those AI-generated pages instead of the organization's website. The system calculates a landing page score for the existing page. If the score exceeds a threshold indicating the existing page is not a good match for user intent, the system triggers AI generation to create a tailored page using the organization's content and the user's context. The updated search results page then shows a navigation link to the AI-generated page instead of the standard URL.

Roger Montti notes that the patent is specifically focused on shopping and advertising use cases, not editorial content. Every example in the patent references e-commerce sites, product listing pages, conversion rates, call-to-action buttons and sponsored content placement. There are no examples related to news sites, blogs or informational content.

Bing Adds GEO to Official Guidelines, Expands AI Abuse Definitions

Matt G. Southern reports that Microsoft rewrote the Bing Webmaster Guidelines to cover both traditional search results and Copilot's AI-generated answers, formally treating grounding results and citations as additional eligibility outcomes. The new guidance recommends stating facts directly, using clear entity names, keeping each URL focused on a single topic, and placing essential information near the top of the page to improve grounding eligibility. NOARCHIVE now blocks use in Copilot responses and grounding results, NOCACHE limits Copilot to URL, title and snippet only, and DATA-NOSNIPPET or NOSNIPPET can reduce citation quality. Bing also expanded its abuse definitions to cover prompt injection and "artificially engineered language" designed to trigger citations or AI responses, not just traditional keyword stuffing.

Straight From the (AI) Source: Is AEO/GEO different than SEO?

Glenn Gabe compiles statements from Google, Microsoft and Perplexity to answer the AEO/GEO vs. SEO question using direct quotes, interviews and platform documentation. The consistent message is that AI search visibility is mostly built on traditional SEO foundations, with clearer structure and formatting helping AI systems interpret content more confidently. Some of the messages:

• Google’s Jeff Dean said LLM-based search is “not that dissimilar” from classic ranking and retrieval, with systems narrowing huge document sets down through multiple refinement stages.

• John Mueller and Danny Sullivan said on Search Off the Record that AI search optimization is essentially a subset of SEO, not a separate discipline.

• Danny also warned against over-chunking content for LLMs, saying Google engineers “really don’t” want site owners turning content into unnatural bite-sized fragments.

• Microsoft’s Krishna Madhavan said SEO fundamentals like structure, freshness, crawlability and trust still matter, and he recommended schema, headings and modular layouts to make content easier for AI systems to parse.

• Microsoft Advertising’s AI search guide says AI systems perform real-time web searches throughout the shopping journey, which keeps traditional SEO essential for being discovered and recommended.

• The same Microsoft guide lists several visibility killers: long walls of text, important answers hidden in tabs, core information locked in PDFs, and key information only available in images.

Is AI Mode Now Dominated by Google's Self-Referencing Links?

Yulia Deda shares the results after analyz 1,321,398 citations across 68,313 keywords in 20 niches and found that Google.com is now the top cited domain in AI Mode responses by a wide margin.

• Google.com accounts for 17.42% of all AI Mode citations, more than the next six domains combined: YouTube, Facebook, Reddit, Amazon, Indeed and Zillow.

• Google's self-citation rate in AI Mode has tripled from 5.7% in June 2025 to 17.42% today.

• 59% of Google citations in AI Mode now point to Google SERPs, while 36.1% point to Google Business Profiles.

• Google ranks as the top-cited source in 19 of the 20 niches analyzed, including Travel at 53.18%, Entertainment and Hobbies at 48.74%, and Real Estate at 30.54%.

• Career and Jobs is the only exception, where Indeed dominates and is cited 3.1 times more than Google.

• The shift means AI Mode is increasingly sending users deeper into Google's own ecosystem rather than directly to publisher websites.

My Take: Google keeping users in the family 🔐 

SEO + GEO Ripples

• Google updated AI Mode recipe results to send more traffic after feedback about "Frankenstein recipes" that synthesized multiple sources without attribution. Now when users search for meal ideas, they can tap on dishes to see links to recipe sites plus a short overview. The rollout is complete, though the UI is not immediately obvious. Recipe images look decorative rather than clickable, which may limit the traffic benefit until users learn the interaction pattern.

• An impostor site is outranking the real NanoClaw project despite 18K GitHub stars. Gavriel Cohen built the open source AI agent platform in February and someone scraped his GitHub README to create nanoclaw.net before he launched nanoclaw.dev. Standard SEO steps including structured data, Search Console submission and takedown notices have not fixed the issue. The same problem appears across DuckDuckGo, Kagi, Bing, Brave and Ecosia, with Mojeek the only engine that got it right.

• ChatGPT sources 83% of its product carousel items from Google Shopping according to a study of 43,000 carousel products. Only 11% matched Bing Shopping results, and just 70 products (0.16%) were found exclusively on Bing. Shopping query fan-outs are distinct from search fan-outs 98.3% of the time, average just 7 words versus 12 for search, and 60% of strong matches come from Google's top 10 shopping positions.

• Google removed its JavaScript accessibility guidance from the official docs after calling the section outdated. The old documentation advised testing sites with JavaScript disabled and using text-only browsers to identify content Google might miss. Google now says it has been rendering JavaScript for years and loading content via JavaScript does not make it harder for search. Most assistive technologies also work with JavaScript now.

• Google updated its image SEO best practices and Discover documentation to clarify how it picks thumbnail images for Search and Discover. Google now explicitly uses both schema.org markup and og:image meta tags as sources. To influence image selection, specify primaryImageOfPage in schema, attach image properties to the main entity, or use og:image. Best practices include choosing representative images, avoiding logos or images with text overlay, and using high resolution with standard aspect ratios. For Discover, Google recommends images at least 1200px wide, at least 300KB file size, and 16x9 aspect ratio.

AI

Labor Market Impacts of AI: A New Measure and Early Evidence

Anthropic introduces a new measure called "observed exposure" that combines theoretical LLM capability with real-world usage data, weighting automated and work-related uses more heavily. The research finds AI is far from reaching its theoretical capability, with actual coverage remaining a fraction of what is feasible.

• Computer Programmers have the highest exposure at 75% task coverage, followed by Customer Service Representatives and Data Entry Keyers at 67%.

• AI currently covers just 33% of tasks in Computer and Math occupations despite theoretical capability of 94%.

• For every 10 percentage point increase in coverage, BLS growth projections drop by 0.6 percentage points, suggesting labor analysts expect slower growth in more exposed occupations.

• Workers in the most exposed professions are more likely to be older, female, more educated, and higher-paid. Those with graduate degrees are 17.4% of the most exposed group versus 4.5% of the unexposed group.

• The research finds no systematic increase in unemployment for highly exposed workers since late 2022.

• Hiring of workers aged 22 to 25 into exposed occupations has slowed by about 14% compared to 2022

• 30% of workers have zero coverage, including Cooks, Motorcycle Mechanics, Lifeguards, Bartenders, and Dishwashers.

• If all workers in the top 10% of exposure were laid off, aggregate unemployment would rise from 4% to13%.

The Ultimate Prompting Guide for Nano Banana

Google published a comprehensive prompting guide for Nano Banana 2 and Nano Banana Pro. The guide covers five prompting frameworks: text-to-image generation without references, multimodal generation with up to 14 reference images, conversational editing with semantic masking, real-time web search integration for grounded visuals, and text rendering in over 10 languages.

Copilot Tasks: From Answers to Actions

Microsoft introduced Copilot Tasks, research preview that shifts AI from conversation to autonomous action. Tasks works in the background with its own computer and browser to complete to-do items like monitoring rental listings and booking showings, compiling daily briefings from email and calendar, drafting replies to urgent messages, and canceling unused subscriptions. Users describe what they need in natural language and Copilot plans and executes across apps and services. Tasks asks for consent before meaningful actions like spending money or sending messages.

AI-Generated Art Can't Be Copyrighted After Supreme Court Declines to Review the Rule

Emma Roth reports that the US Supreme Court declined to hear Stephen Thaler's case over whether AI-generated art can obtain copyright protection. The decision leaves in place earlier rulings that found AI-created work without human authorship is ineligible for copyright. The case began after the US Copyright Office rejected Thaler's 2019 attempt to register an image created by his algorithm, A Recent Entrance to Paradise, and reaffirmed in 2022 that the work lacked the required human authorship. A district court judge ruled in 2023 that human authorship is a bedrock requirement of copyright, and a federal appeals court upheld that decision in 2025. The ruling also aligns with the Copyright Office's recent guidance that AI-generated artwork based only on text prompts is not protected.

Google's New Command-Line Tool Can Plug OpenClaw Into Your Workspace Data

Ryan Whitwam reports that Google released a new Google Workspace CLI project gws that bundles Workspace APIs into a command-line tool designed for both humans and AI agents, with built-in support for OpenClaw. The tool covers Gmail, Drive, Calendar and other Workspace products, outputs structured JSON, and includes more than 40 agent skills. Google positions it as a cleaner alternative to more complex MCP setups. The catch is that Google explicitly says this is “not an officially supported Google product,” functionality may change dramatically, and users are on their own if workflows break.

My Take: It sounds cool, but I’m still using Google’s API to access the various tools. This is still early and most likely will have issues. OpenClaw can be fragile itself, so adding another fragile layer can lead to more issues that I don’t want to deal with.

Claude Marketplace

Anthropic launched Claude Marketplace in limited preview, giving enterprise customers a way to spend part of their existing Anthropic commitment on third-party Claude-powered tools. Instead of negotiating separate budgets and invoices for each partner product, companies can apply existing Anthropic commitments across approved tools. Anthropic is launching with GitLab, Harvey, Lovable, Replit, Rogo and Snowflake, with billing handled directly through Anthropic.

AI Ripples

• OpenAI released GPT-5.3 Instant, an update to ChatGPT's most-used model that reduces unnecessary refusals, tones down overly cautious or preachy responses, and improves web search contextualization. The company reports hallucination rates dropped 26.8% with web use and 19.7% without web access on higher-stakes evaluations. The update also cuts back on phrases like "Stop. Take a breath" and aims for a more natural conversational style.

• OpenAI released GPT-5.4, its most capable frontier model for professional work with native computer-use capabilities and support for up to 1M tokens of context. It’s one of the quickest models out there at this level. GPT-5.4 is priced at $2.50 per million input tokens and $15.00 per million output tokens. GPT-5.4 Pro is also available for maximum performance on complex tasks.

• Google released Gemini 3.1 Flash-Lite, its fastest and most cost-efficient Gemini 3 series model at $0.25 per million input tokens and $1.50 per million output tokens. It outperforms 2.5 Flash with 2.5X faster time to first token and 45% higher output speed, achieving 1432 Elo on Arena.ai and 86.9% on GPQA Diamond.

• OpenAI launched ChatGPT for Excel in beta, an add-in powered by GPT-5.4 that builds, analyzes and updates spreadsheet models directly inside workbooks. It supports financial modeling, scenario analysis, and tracing formula dependencies while preserving Excel-native structure. On OpenAI's internal investment banking benchmark, performance improved from 43.7% with GPT-5 to 87.3% with GPT-5.4 Thinking.

• Google added Canvas to AI Mode in Search, a dedicated space for drafting documents and building interactive tools. Users can describe a dashboard or application and get a working prototype that pulls live data from the web and Google's Knowledge Graph. Canvas is available in the U.S. in English.

• Grammarly's Expert Review feature uses journalist identities without permission, including Verge editors and tech writers from Wired, Bloomberg and The New York Times. The feature claims to offer feedback "inspired by" experts but did not notify or seek consent from those named. Sources linked from suggestions often pointed to spammy archive copies or unrelated pages, suggesting the AI may be basing suggestions on different people's work than the name attached.

• OpenAI scaled back ChatGPT Instant Checkout plans, moving purchases from native in-app checkout to connected retailer apps instead. Users research products in ChatGPT but rarely complete purchases there, and only a small number of merchants were actively using the feature.

MARKETING

Why Most Video Ads Fail (And What Metrics Actually Matter)

Greg Jarboe says that most video ads fail because marketers optimize for platform outputs like views, impressions, and completion rates instead of business outcomes like attention, persuasion, and action. He says the real problem is not usually targeting, budget, or platform choice. It is weak strategy. Ads are still made like old TV commercials even though modern platforms reward relevance, retention, and native creative. He highlights several recurring issues: brands waste the opening seconds on logos and branding cues, overvalue polished production that feels like an ad, separate creative from distribution context, and run tests without clear hypotheses.

How AI Tools Influence the Modern Buyer Journey

Margarita Loktionova presents Semrush’s survey of 1,030 U.S. consumers who have tried AI tools and found that AI is already embedded across the shopping journey, from early discovery to final comparison. Some of the takeaways:

• 48% use AI daily, 85% use it at least weekly, and 55% use it for product research at least weekly

• 77% use AI and traditional search together

• Google is the more common starting point at 33% versus AI at 26%.

• 43% have discovered a new brand through AI

• 50% have made a purchase after using AI during research

• 69% expect AI to play a bigger role in how they shop in the future

• Only 20% say a brand stands out because it appears earlier in the response, while clearer explanations, value context, fit for needs, and direct comparisons have more influence.

Why AI-Built Websites All Look the Same

YC’s Aaron Epstein and Raphael Schaad review startup websites built with AI coding tools. They found that LLMs reproduce patterns from training data, so purple gradients, scroll fade-ins and flashy hover effects now appear everywhere without adding brand identity. They critique animations that distract from the product, including cursor-tracking effects, buttons that move toward the user, and scroll hijacking that breaks navigation. The reviewers recommend treating AI like a junior designer and editing aggressively rather than accepting every suggestion.

CONTENT

How to Chunk Content and When It's Worth It

Tania Brown responds to Google's recent criticism of chunking as an AI-optimization hack, arguing that chunking helps readers scan and AI systems retrieve passages accurately. The technique organizes text into self-contained paragraphs where each covers one complete idea. A reader should grasp the concept from a single paragraph without hunting for context. The author acknowledges that not every piece needs chunking. Skip content that already performs well with AI citations, is scheduled for comprehensive rewrites, or relies on narrative flow and emotional arcs like thought leadership essays and brand storytelling.

TOOLS AND RESOURCES

How to Turn Claude Code Into Your SEO Command Center

Will Scott walks through setting up Claude Code to pull live data from Google Search Console, GA4 and Google Ads via Python scripts, then ask cross-source questions without spreadsheets. He says the setup takes about an hour and enables queries like "which keywords am I paying for that I already rank for organically?" A paid-organic gap analysis for a higher education client identified 2,742 search terms with wasted ad spend, 351 opportunities to reduce paid spend where organic was strong, and 41 content gaps where paid was the only presence.

WAYS WE CAN WORK TOGETHER

Floyi - The only AI-powered tool that builds 4-level topical maps. Don’t just plan your content strategy - make it unstoppable.

TopicalMap.com Service - Let us do the heavy lifting. We handle the research, structure, and strategy. You get a custom topical map designed to boost authority and dominate your niche and industry.

Topical Maps Unlocked 2.0 - Unlock the blueprint to ranking success. Master the art of structuring content that search engines (and your audience) love - and watch your rankings soar.

AD

The Headlines Traders Need Before the Bell

Tired of missing the trades that actually move?

In under five minutes, Elite Trade Club delivers the top stories, market-moving headlines, and stocks to watch — before the open.

Join 200K+ traders who start with a plan, not a scroll.

Get the Next Alert