🏛️ Official Updates
3 new ways to plan and book travel in Search
Google just announced 3 new ways to plan and book travel directly in AI Mode. This official update matters for GEO/SEO practitioners tracking how AI-powered search reshapes travel discovery and conversion.
First, AI Mode now includes Google Flights’ price tracking. You can set fare alerts mid-conversation across 300+ airlines in 180+ countries. Second, AI Mode shows points and miles costs for flights and hotels. Initial partners include American Airlines, Hilton, and Wyndham, with Hyatt and Lufthansa joining soon. Third, hotel booking happens entirely in AI Mode. Partners like Booking.com, Expedia, and Marriott let users complete reservations via Google Pay.
I recommend reading the full post to understand the booking flow specifics. The merchant-of-record model and partner list reveal how Google is monetizing AI-driven travel intent. The post focuses on product features rather than strategic analysis. Still, the shift is clear: AI Mode moves from recommendation to transaction. Travel brands should prepare for AI Mode as a direct conversion channel.
5 ways to upgrade your home decor with Google Search
This article gives SEO practitioners a useful window into Google’s evolving visual and AI search features — but treat it as product marketing, not technical guidance.
I think that’s the right frame for the “5 ways to upgrade” home decor angle. The piece showcases five tools: AI Mode for furniture visualization, Lens for identifying vintage finds, Circle to Search for shopping from social feeds, Search Live for DIY video guidance, and price history tracking. Notably, “home decor inspo” searches jumped 300% in the past month. “Fish wallpaper” rose 140%. Vintage rugs and wood coffee tables topped furnishing trends. For SEO, the takeaway is clear: image optimization, structured data, and visual search relevance are becoming non-negotiable. I recommend reading this as a trend tracker. The score is 7/10 — directionally valuable, but thin on implementation details.
Introducing the Admin plugin for ChatGPT Work and Codex
I think the Admin plugin announcement deserves attention even if it won’t change your SEO workflow tomorrow. OpenAI shipped an Admin plugin for ChatGPT Work and Codex. It lets workspace admins handle usage analytics, member management, and permission changes in one conversation.
Three details stand out. Admins can review credit usage, adjust limits, and approve spending requests without switching tools. The plugin routes pending requests to Slack or Microsoft Teams for approval. OpenAI’s IT team resolved about 45% of ticket volume with ChatGPT Work agents.
I recommend scanning this if you manage AI tool adoption. The permission-aware design preserves existing roles and controls. Support volume roughly doubled, yet the team cleared its backlog.
Direct SEO value stays thin. The real signal is how OpenAI pushes AI agents into operational workflows. That trend will reach content operations and client reporting. Understanding these admin capabilities helps you plan your AI tooling strategy.
🤖 GEO·SEO Highlights
What Is WebMCP? How to Prepare Your Website to Serve AI Agents
WebMCP is the proposed Google-Microsoft standard for preparing websites to serve AI agents. This Moz guide delivers the clearest implementation playbook I have seen. Matt Hollingshead explains how to expose structured tools like “search products” or “book a table” through HTML attributes or a few lines of JavaScript.
ChatGPT Rebuilt Its Search Tool, I Read The New Language It Speaks
OpenAI quietly swapped ChatGPT’s search backend in four days, and this technical teardown by Suganthan Mohanadasan shows exactly what changed. JSON fan-out calls became a compact pipe-delimited query language. He decoded the new format by capturing his own browser traffic. He ran eight test questions on a single ChatGPT Plus account, spanning commercial, local, news, and finance.
The new tool call format is remarkably readable. Each line is one search with pipe-separated fields: call type, query, number, and optionally a domain. The third field is a freshness window matched to how fast answers go stale. Stock prices get a 2-day window, football results get 7, commercial research gets 30, and earnings guidance gets 90. For GEO practitioners, this is gold. Pricing or comparison pages older than 30 days sit outside the default window. This happens exactly on the queries where brands get compared.
The format also introduces call types. fast is the web search. product is a catalogue lookup for physical goods, checking product cards and merchant offers. The domain slot confirms ChatGPT visits sites it already knows. The old search_queries metadata field is gone entirely.
I recommend this piece if you optimize for AI search discovery. The freshness data alone justifies the read. Update key pages inside a 30-day cycle and put answer sentences where excerpts begin.
AI Brand Preference Now Splits By Generation, Claude Leads Gen Z More Than 7-To-1
YouGov’s newest AI brand preference data exposes the steepest generational split I’ve seen in any AI ranking. Greg Jarboe tells you why momentum metrics never equal trust. Search Engine Journal breaks down YouGov’s second U.S. AI brand rankings by generation, and the gaps are dramatic.
Claude leads Gen Z AI users at 10.6% preference versus 1.4% among Boomers plus. That is a 7.6-to-1 divide. ChatGPT wins every age group, from 44.4% among Gen Z down to 24.5% among Boomers plus. Gemini performs best among Gen X at 22.8%. Copilot moves in the opposite direction, hitting 12.3% among Boomers plus and only 4.4% among Gen Z. Alexa posts 11.0% among Boomers plus and a clean 0% among Gen Z.
I find the methodology point most valuable. YouGov applied the identical measurement framework to AI assistants and fast-casual restaurants. Chicken Salad Chick tops the emerging dining brands list on Consideration, buzz, and recommendation. No one mistakes a regional chicken chain’s familiarity for proof that its food is safe. Yet we routinely read AI Consideration as evidence that users trust the answers.
My recommendation stands. Separate trust signals from momentum signals in your GEO work. Brand preference reveals which tool entered a generation’s routine first. Real trust requires proof, especially since only 28% of Americans trust AI assistants. Build content that demonstrates accuracy instead of chasing buzz alone.
Reports Indicate Google’s Spam Update Focused On SEO AI Content
Reports indicate Google’s August spam update targets mass-generated SEO AI content. This Search Engine Journal piece compiles the strongest community evidence yet. I recommend it for anyone scaling content operations.
The core insight is the Scalable Cluster Termination System, or S-CTS. Google designed it to terminate entire networks of AI spam. Japanese SEOs report that fully automated sites dropped across the board. Sites with early manual posting survive because they accumulated user trust signals.
I think the most valuable takeaway is that production method matters more than AI usage. Mass production for ranking manipulation draws penalties. Human-reviewed AI content remains viable.
One black hat forum member called AI slop the new doorway page. That framing helps. I recommend auditing your content pipeline now. Focus on unique value over volume. Automated keyword-driven content carries real risk.
Google Brings Developing-Topic Link Carousels To AI Mode
Google brings developing-topic link carousels to AI Mode, giving timely articles a prominent spot inside AI answers. I think this is a practical AI Search update for publishers and SEOs.
The carousel shows article cards with an image, headline, source, and date. Robby Stein, VP of Product for Google Search, announced the rollout on X. The format launched in AI Overviews in May and now covers both generative Search surfaces. Interestingly, Preferred Sources get highlighted inside the carousel. Google reports 600,000+ unique Preferred Sources, up from 345,000 in May.
Google only shows the carousel for some queries, so visibility isn’t guaranteed. AI Mode links generate Search Console impressions and clicks, but no filter isolates the carousel. I recommend tracking AI Mode performance for trending topics and watching for these placements.
Your Promo Code Field May Be Sending Revenue To Affiliates
The promo code field in your checkout is quietly sending revenue to coupon affiliates. This article by Loren Baker exposes how a blank discount box creates a branded search moment at peak purchase intent. It delivers a clear strategy for reclaiming that traffic and stopping avoidable commission leakage.
I think this is essential reading for ecommerce SEO and digital marketing teams. The article reframes the problem as a multi-channel revenue leak that reaches far beyond rankings. Baker connects checkout UX to attribution, affiliate costs, margins, and conversion rates.
The economics are compelling. On a $900 order, a 10%-20% commission hands $90-$180 to a partner who never introduced the customer. The brand’s own organic, paid, email, or direct channels lose credit for the sale. If the shopper applies a discount code, the retailer takes two reductions on a single order.
Small search volume still carries high commercial intent. A midsize brand saw roughly 760 monthly searches for coupon-related terms. Those searchers have credit cards and maximum purchase intent. Baker argues that 760 buyers beat tens of thousands of passive visitors.
The strategic fix is a brand-owned savings page. This destination names the source, answers the shopper’s fear of missing a deal, and states the current offer. It can feature a promotion, seasonal offer, email incentive, loyalty benefit, or an honest statement that no public code exists. Accuracy and connection to the purchase journey matter most.
I recommend reading the full SEJ Pro presentation if you oversee ecommerce revenue or affiliate partnerships. Measure success through assisted revenue and checkout behavior. This article gives you the commercial argument to win budget beyond the SEO team.
Semantic Relevance Isn’t One Metric — It Has Three Jobs
Semantic relevance isn’t a single score you can optimize in isolation. That’s the core argument from Lumar’s latest GEO deep-dive, and I think it’s one of the most practical frameworks I’ve seen this year.
The article breaks semantic relevance into three distinct stages: candidate selection (page-level topical fit), chunk retrieval (passage-level answer match), and citation (self-contained fact quality). Each stage demands different content treatment. A page can nail one stage and fail the next — which explains why some highly relevant pages rarely get cited by AI systems.
I recommend this piece for anyone doing GEO work. It gives you a diagnostic lens: clear page topics support candidate selection, focused answers support chunk retrieval, and self-contained passages support citation. Lumar even includes a summary table mapping each stage to what it optimizes, which makes the framework easy to apply.
My one critique: concrete examples of pages failing at each stage would strengthen the argument. Still, the framework alone is worth the read for separating genuine relevance from surface-level similarity.
What’s a Good GA4 Engagement Rate? 500+ Websites, Analyzed
Stop guessing what a good GA4 engagement rate is. Siege Media analyzed 16 billion sessions across 500+ websites and delivered the definitive benchmark. The median engagement rate is 51%. I now treat 50% as the baseline for “good” and 55% as strong.
This study stands out because it breaks data down by industry, device, and channel. E-commerce and retail lead at 54.4%. Health and wellness follows at 51.6%. SaaS and B2B trails at 46.6%, since much of that traffic lands on top-of-funnel content during long research cycles.
Device mix matters more than most people realize. Desktop sessions engage at 55.3%. Mobile sits at 48.8%. The seven-point gap means your device split alone can shift your overall rate. I recommend checking that breakdown before assuming your content underperforms.
The channel data is the real goldmine for GEO practitioners. Organic search drives a 63.4% median engagement rate. AI assistants follow at 60.3%. That tells me AI-referred traffic carries genuine intent. Paid search sits at 60.4%. Paid social trails at just 20.2%, because interruption-based traffic engages poorly regardless of page quality.
I recommend comparing your site against these segments rather than chasing a generic global average. The article also includes a prioritized list of levers to improve engagement. This is the reference I will use for all future GA4 benchmark questions.
AI Search Didn’t Remove Cognitive Load, It Moved It
AI search didn’t eliminate cognitive load — it moved it.
AI search has moved the cognitive burden from retrieval to verification, and this shift changes what SEO content must survive. Duane Forrester’s analysis, rooted in Jakob Nielsen’s cognitive load budget theory, argues consumers now audit pre-assembled answers rather than build their own. The evidence is concrete. A 2026 ACL study shows that generative search synthesizes before exposing sources. Microsoft’s analysis of 200,000 Bing Copilot conversations confirms users keep the goal while delegating the information work. The sharpest finding comes from Li and Aral’s experiment. Citations increased trust even when references were hallucinated, and trusted results received less scrutiny. That inverts search’s old bargain. Synthesis now leads; evidence trails. I recommend this piece for SEOs still optimizing for ranked lists. Your content must survive an invisible synthesis process, not just a human click. The verification burden has moved to the user’s side of the screen, so winning means building answers that hold up under audit, not just ones that rank.
OpenAI Adds WebMCP Site Tools To ChatGPT’s Browser
OpenAI adds WebMCP to ChatGPT’s desktop browser, and this changes how sites serve AI agents. I think this is the most important agent-site update for SEOs this year.
The article reports that WebMCP lets a webpage register JavaScript functions as tools. Agents can search documents, edit files, compare travel options, or update shopping carts. This differs from server-based MCP; the page itself provides the tools. Tools stay tied to the current page and signed-in session.
I recommend reading the full piece before you build your agent strategy. The feature requires GPT-5.6 Sol or Terra. Luna has WebMCP disabled, and Enterprise or Edu workspaces lack access. Developers can test in Chrome today via an origin trial. OpenAI also flags prompt injection and data exfiltration risks, so permission flows matter. Interestingly, the documentation stops short of explaining rankings or discoverability.
My take: WebMCP gives site owners a structured way to define agent actions. Google already advises building for AI agents; this move makes that concrete. I recommend testing it now, because early adopters will shape how ChatGPT surfaces site tools.
How To Advertise In Google AI Mode For Ecommerce
If you want to advertise in Google AI Mode for ecommerce, fix your product feed first. That is the core argument of this Search Engine Journal piece by Tony Adam. Google builds AI Mode ads from Merchant Center data, not advertiser copy. Bidding matters less than product data quality.
How To Become A Strategic Partner & Keep Clients Years Longer
This Search Engine Journal article shows SEO agencies exactly how to become a strategic partner and keep clients for years longer. Sophie Brannon of StudioHawk US makes the case with fresh 2025 ANA/4As tenure data.
The numbers are decisive. Client-agency relationships now last seven years, more than double the 3.2-year average from 2016. Clients running competitive reviews stay only 3.8 years versus 8.1 years without them. Full-service agencies hold clients for 7.3 years, while media-only firms manage 3.7 years.
Brannon pinpoints three partner habits. Partners flag bad news first, with an explanation and plan attached. Partners connect SEO reporting to the pipeline or bookings that leadership actually reviews. Partners share opinions beyond the contract scope, even when it touches conversion optimization.
I recommend the “bonus metric” tactic most of all. Ask your point of contact which targets drive their annual bonus, then build reporting around those numbers. Supporting your client’s personal success makes your retainer the last line item anyone cuts. I have run client relationships for a decade, and this advice aligns with what keeps retainers intact.
Interestingly, a Databox and ZenPilot survey of 300 agencies found communication drives client satisfaction more than delivered results. Build a communication cadence that clients rely on. That single shift moves you from order-taker to trusted advisor.
Judge Says Google/Publisher AI Situation “Seems Really Unfair”
Judge says Google’s AI Overviews arrangement “seems really unfair” — and that observation could reshape publisher-SEO economics. This is the most important antitrust signal for search professionals since the 2024 monopoly ruling.
Judge Amit Mehta is the same judge who declared Google a monopoly. On Tuesday, he heard Penske Media’s motion in its antitrust lawsuit over AI Overviews. Jason Kint, CEO of Digital Content Next, attended and tweeted the blow-by-blow.
Google’s attorneys called AI Overviews a “product improvement.” They described the historic crawl-for-traffic bargain with publishers as a “vague historical course of dealing.” Mehta pushed back hard. He said the situation “all seems really unfair.” He noted that publishers lack control over their content. He said Google builds AI “on the backs of the publishers.” And he made a critical point: product improvements remain subject to antitrust scrutiny.
I find the market argument most compelling. Mehta observed that a market is forming for GenAI inputs. OpenAI and Perplexity pay millions for content. Google uses its search monopoly to obtain the same inputs at zero cost. That raises rivals’ costs and barriers to entry.
Publishers can technically opt out of AI training and AI Overviews. That choice costs publishers the few remaining traffic crumbs. I recommend that every SEO and publisher read this coverage. The dismissal motion remains pending before Mehta. But his questions suggest Penske’s theory is receiving serious consideration. If the case survives, the publisher-Google relationship faces fundamental change. Our traffic strategies depend on the outcome.
Why International SEO Needs Machine-Recognizable E-E-A-T
International SEO needs machine-recognizable E-E-A-T, and Motoko Hunt’s SEJ deep-dive explains why. The article delivers a sharp warning: AI flattens localized expertise across 40 regional sites into one global brand impression. Humans see distinct local credibility. Models see one generic brand.
Hunt names this the credential gap. US-trained LLMs struggle to recognize professional certifications, titles, and licensing systems from other markets. We can publish flawless local content, and the model still misses the expertise. Source-of-truth status answers who the company is. E-E-A-T answers whether it knows the subject. AI evaluates those claims separately.
Hunt proposes solving for both audiences: we demonstrate expertise for human readers, then make it legible to machines. She recommends explicit, machine-readable market boundaries and structured credential evidence. My take: this reframes international SEO completely—brands earn authority market by market, in forms AI can learn.
Peec AI alternatives for AI visibility monitoring in 2026
I recommend this HubSpot guide to Peec AI alternatives because it moves beyond dashboards into revenue attribution. The article ranks 10 AI visibility platforms for 2026. HubSpot AEO leads for CRM-native attribution, Writesonic GEO for end-to-end optimization, and Profound for enterprise analytics.
I appreciate the practical buyer framework. The evaluation covers 10 AI models — ChatGPT, Perplexity, Gemini, Grok, Claude, and Copilot. Citation analysis reveals which sources drive AI recommendations. The RFP checklist gives teams a shared evaluation language before trials start.
The strongest section connects AI visibility to the CRM pipeline. The activation plan and pricing breakdown provide immediate next steps for GEO teams. I recommend pairing this article with an internal audit of your monitoring stack. The comparison table alone justifies the read.
MBA Metrics for SEO: How to Speak Leadership’s Language
Boardroom decisions run on market share and revenue rather than keyword rankings. I recommend this Ahrefs guide on MBA Metrics for SEO because it builds a translation layer between SEO work and leadership priorities. Market share maps to Share of Voice, revenue growth maps to organic traffic value, and brand demand maps to branded search.
The evidence is compelling. James Hankins analyzed 30 case studies across 12 categories and seven countries. Share of search drives roughly 83% of market share. Les Binet’s earlier research confirms this correlation.
I especially value the Share of Voice reframe. Ask ‘how much of the market are we winning?’ rather than ‘are we ranking?’ Ahrefs Rank Tracker computes it automatically. Brand Radar tracks the AI equivalent in ChatGPT and Gemini answers.
My take: use this framework to earn a seat at the strategy table. Leadership responds when SEO speaks their language.
The 15 Best SEO Books Backed by Our SEO/GEO Team [2026]
This ranking of the 15 best SEO books comes from Siege Media’s own practitioners. The team manages over $148 million in monthly client content value, so these picks carry real operational weight. I think the list stands out because it places GEO books first, treating AI search as today’s reality rather than a future trend.
The selections split into four clear categories: GEO and AI search, beginner fundamentals, leadership strategy, and reference volumes. Notable picks include Ross Hudgens’ ‘GEO,’ Peter Kent’s ‘SEO for Dummies,’ Eli Schwartz’s ‘Product-Led SEO,’ and the updated ‘The Art of SEO.’ Each entry lists the author, publish date, page count, and recommendation reasons. The authors advise treating books as foundations for thinking, not line-by-line scripts, since search changes faster than publication cycles.
I recommend this article for anyone building a team reading list. The methodology is transparent. The authors cross-referenced expert roundups, community threads, and review sites, then vetted results against their internal hiring recommendations. The 2026 publication dates keep AI search guidance current. Technical depth is limited, but as a curated starting point, this delivers more practical value than recycled listicles.
Analysis: Reddit Appears in 83.9% of Google’s Discussions and Forums Results
This Ahrefs analysis finds Reddit appears in 83.9% of Google’s Discussions and Forums results worldwide. I recommend it to SEO practitioners tracking forum content strategies. The study quantifies Reddit’s SERP dominance with hard data.
Reddit leads 87.8% of US discussions results. Average SERPs carry 1.29 Reddit links. Quora trails in second place. Google’s preference for Reddit is now measurable.
The strategic takeaway is direct. Building a genuine presence on Reddit outperforms launching your own forum. The data supports reallocating resources toward community participation. The study also maps which query categories favor Reddit most. That helps prioritize engagement topics.
Specific, current, and actionable. This earns a solid 8/10 from me.
Reddit’s citations in ChatGPT fall from 3.8% to 0.5%
Reddit’s citations in ChatGPT fell 86% in four days.
This Semrush analysis delivers the data SEO teams need right now. Promptwatch tracked Reddit’s steady 3.8% share of ChatGPT citations from July 18 through August 7. Then came the collapse to 0.5% between August 14 and 17. OpenAI denies changing its source selection. The timing aligns with an August 8 change in ChatGPT Search routing. I recommend treating this as an early signal worth watching. Reddit citations dropped only 11% in Google’s AI Overviews and 31% in AI Mode. That concentration points to ChatGPT-specific mechanics. Reddit still offers direct customer sentiment for research. I suggest tracking where AI citation traffic moves using Semrush’s AI Traffic Dashboard and Competitor Research tools. This article earns an 8/10 for making complex visibility data actionable.
AI Tools Recommend Brands But Cite Other Sites, Data Shows
This data gives ecommerce brands a clear picture of AI citation behavior.
Shero Commerce analyzed 1,851 citations from Google AI Mode, ChatGPT, and Perplexity. The headline finding: AI tools recommend brands but cite other websites. Brand-owned pages earned only 2.8% of citations. Third-party review sites captured 59% of source links. In 159 brand recommendations, the brand’s own page appeared just 31% of the time. Google AI Mode cited brands in 9.5% of relevant shopping queries. The report also found 20% of Shopify product descriptions contained duplicate content. Duplicate text complicates AI attribution. Original, distinct product copy gives brands a stronger path to citations. The report lacks testing on whether rewrites change source selection. I recommend treating these numbers as a baseline for your AI visibility audits.
