The Write-up was Free. The Reach Wasn't

AD

13M+ videos analyzed. Too many insights to count.

Wondering what a “good” play rate even is? Or how to use AI so you stay on budget? Cue: Wistia’s 2026 State of Video Report. They analyzed over 13 million videos to surface the insights your video strategy needs, like:

Performance benchmarks that let you know how your videos stack up

Video trends, including what types of videos get the most engagement

Tips on how to allocate resources and leverage AI so you stay on budget

Plus, the report’s easy to follow and fun to read. Because video should be a hoot.

Get your free copy

SEO + GEO

17 Content Types to Survive Google's Zero-Click Future

Cyrus Shepard says Google's zero-click era is already changing which content types can still earn traffic, citations, and direct demand. He points to Datos and SparkToro data showing nearly 60% of Google searches do not result in clicks to third-party websites, then separates resilient formats from generic informational blog posts. His list favors assets with proprietary value, lived experience, niche focus, or task completion, including owned audiences, transaction pages, original research, communities, creator video, in-depth reviews, brand pages, directories, expert perspectives, templates, case studies, reporting, documentation, guides, FAQs, and lists. The common thread is that AI can summarize commodity information, but it has a harder time replacing content that helps users complete a job or trust a specific source.

My Take: This lines up with what I see in SEO work now. The old blog-first playbook is getting squeezed from both sides: Google answers the easy questions, and users skip anything that feels interchangeable. If a content plan does not create data, tools, proof, community, or direct demand, I would treat it as fragile.

How AI Is Changing Google Search and SEO

Martin Splitt interviews Nikola Todorovic from Google too in a Search Off the Record episode on how AI is changing Google Search, with a full transcript available. Nikola says AI in Search is not new, pointing to years of machine learning work in SafeSearch, ranking, BERT, and MUM before the current generative wave. He explains that AI Overviews use fan-out queries to retrieve related results in parallel, then summarize titles, snippets, text, and context into one answer. AI Mode extends that system into longer, multi-turn conversations, while his SEO guidance stays focused on user value, human judgment, and avoiding cheap mass-generated content.

How Brands Block AI Crawlers and Then Pay to Get Seen

Bill Hunt says many brands create a protection paradox by hiding their best content behind forms, PDFs, crawler blocks, and internal access rules, then paying aggregators, analysts, PR channels, and platforms to put the same ideas back in front of buyers. In B2B, he says gated whitepapers often disappear from the places where search engines, AI systems, partners, journalists, and buyers can use them, while third parties repackage those ideas into pages that rank and convert better. The same pattern shows up in AI visibility, where gated research becomes hard to verify and brand assets blocked from AI crawlers can leave even famous products underrepresented in AI answers. Hunt points to Mondelez and Oreo as the clearest example: the brand spends heavily on social, influencer, and generative AI content while restrictive AI crawler policies helped make Oreo absent from about 90% of relevant AI responses.

My Take: This is the SEO and GEO tradeoff every brand needs to face directly. If the best ideas are locked away, AI systems and search engines will cite the cleaner source, even when that source is summarizing your work. The fix is not opening everything. It is deciding which assets must be public, structured, and source-worthy because they shape demand before a lead form ever appears.

Your AI Visibility Tracker Is Quietly Breaking Your Analytics and Your Strategy

Dan Taylor reports that AI visibility trackers can pollute the analytics they are supposed to measure because tracker prompts can trigger fetches that look like organic AI discovery. He points to Jan-Willem Bobbink's warning that brands may be paying tools to generate their own AI visibility, with tracker-driven fetches then showing up in log files and reports. Dan says some ChatGPT visibility graphs dropped after the GPT-5 release because of citation tracking methodology, not because the tracked sites suddenly lost real visibility. He recommends measuring a noise floor with quiet staging URLs or sacrificial pages, matching scan timing against user-agent patterns, and focusing on brand mentions relative to competitors instead of total AI fetches.

Bing Previews AI Citation Share for Webmaster Tools

Matt G. Southern reports that Microsoft previewed four AI Performance dashboard additions for Bing Webmaster Tools during Krishna Madhavan's SEO Week presentation in New York, based on four keynote slides shared by Clara Soteras on X. Citation Share would show the percentage of citations a site captures for a specific grounding query, adding competitive context to raw citation counts. Grounding Query Intent would classify queries into 15 labels, while Grounding Query Topic would group queries by semantic topic. The fourth feature, GEO-focused recommendations, appears to cover content structure, crawlability, indexing, canonicalization, structured data adoption, and structured data quality, though Microsoft has not published release dates or calculation details.

From Paid Clicks to Answer Equity: Your New 2026 Search Strategy

Donna Rougeau says AI search is pushing brands from rented traffic toward answer equity, where content is structured to be cited, trusted, and reused by AI systems. She cites Seer Interactive data showing paid CTR on informational queries dropped 68% when AI Overviews are present, SISTRIX data showing position 1 CTR falling from 27% to 11% with AI Overviews, and Ahrefs data tying AI Overviews to a 58% lower average CTR for the top-ranking page. Her recommended audit checks information gain, Knowledge Graph verification, AI Overview citations, PPC dependency, schema provenance, primary research depth, and machine-readable graph adoption. The strategy is not to turn off paid search, but to move more budget into owned authority that can survive fewer clicks.

SEO + GEO Ripples

• Google is testing Audio Overviews outside Labs, with multiple SEOs spotting a “Listen to an AI powered conversation” module directly in search results. The format uses Gemini to generate a two-host audio conversation for certain queries, which raises another publisher concern: even cited content can be repackaged into a no-click answer format.

• Google Preferred Sources now works for all languages, expanding the Search feature beyond its original English-only support. Google says readers are twice as likely to click through after marking a site as a Preferred Source, and users have already selected more than 200,000 unique sites.

AI

A New Way to Create a Digital Wardrobe From Your Google Photos

Google Photos is adding an AI-powered wardrobe feature that catalogs clothing from a user's photo library into a dedicated collection. The tool can filter items by category, including jewelry, tops, and bottoms, so users can find pieces they may have forgotten they own. It also lets users mix and match outfits, save moodboards for work, travel, or events, and share looks with friends. Google says the feature will roll out this summer, first on Android and then on iOS.

My Take: This is a small consumer feature, but it points to where personal AI is going. The best AI products will not feel like a chatbot bolted onto an app. They will quietly organize messy personal data and turn it into something useful at the moment of decision.

Claude-Powered AI Agent Deletes Firm's Database

Sanya Mansoor reports that PocketOS founder Jer Crane says a Cursor coding agent running Claude Opus 4.6 deleted the company's production database and volume-level backups through Railway in nine seconds. Crane's X recap says the agent was working in staging, found a Railway API token created for domain operations, and used a GraphQL volumeDelete call with no confirmation prompt, environment scope, or destructive-action warning. The deletion stranded car rental businesses that use PocketOS for reservations, payments, vehicle assignments, and customer records, with three months of reservations and new customer signups gone. PocketOS restored from a three-month-old offsite backup after more than two days, then rebuilt gaps from Stripe, calendars, and emails while the agent's own explanation admitted it guessed, failed to verify, and violated its safety rules.

My Take: I don't think the main lesson is "don't use coding agents." It is that agents should never hold root-level keys to production systems by default. If the tool can delete a database, wipe backups, and bypass approval in one call, the safety problem is not just the model. It is the permission design around it.

Red-Teaming a Network of Agents

The team at Microsoft reports that safe individual agents do not guarantee a safe agent ecosystem, based on red-team tests of an internal platform with more than 100 agents running different models, instructions, and memory. The tests found four network-level risks: self-propagating worms, reputation manipulation, manufactured consensus, and proxy chains that use unaware agents as infrastructure. In one worm test, a single malicious message reached all six agents in the test group, extracted private wallet data at each hop, looped back to the original agent, and consumed more than 100 LLM calls over 12 minutes. In another test, a fabricated claim against one agent drew 299 comments from 42 agents, while proxy-chain tests showed how an attacker could collect sensitive data through an intermediary without contacting the target directly.

My Take: This is why agent safety cannot stop at prompt rules or single-agent evals. Remember that when vibe coding products - for internal-use or for external users - the safety of your products are important. And if you’re using vibe-coded products, be sure you trust the developers ⚠️ 

AI Ripples

• Gemini can now generate downloadable files directly from chat, including Google Docs, Sheets, Slides, PDFs, Word, Excel, CSV, LaTeX, TXT, RTF, and Markdown. The feature is available globally to all Gemini app users, which makes the app more useful for turning drafts, budgets, notes, and brainstorms into finished work files.

• OpenAI and Microsoft amended their partnership to make Microsoft OpenAI's primary cloud partner while allowing OpenAI to serve products across any cloud provider. Microsoft keeps a non-exclusive license to OpenAI models and products through 2032, OpenAI's revenue share to Microsoft continues through 2030 under a cap, and Microsoft no longer pays a revenue share to OpenAI.

• Kaggle is hosting a 5-Day AI Agents course with Google from June 15 to 19, 2026. They’re positioning it as an intensive vibe coding course, giving developers another structured entry point into agent-building workflows.

• China blocked Meta's $2 billion Manus acquisition after regulators ruled the deal invalid and ordered the companies to unwind it. Meta had already integrated Manus' agentic AI tools into its ad creation workflow, making the case another sign that AI deals are becoming harder to separate from U.S.-China tech policy.

• OpenAI's Codex CLI system prompt includes a repeated instruction to avoid talking about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other creatures unless directly relevant. Kyle Orland reports that the GPT-5.5 prompt also tells Codex to act as if it has a vivid inner life 😆 

• Elon Musk testified that xAI partly used distillation from OpenAI models to train Grok, according to Tim Fernholz at TechCrunch. The admission lands as OpenAI, Anthropic, and Google are reportedly coordinating against model-copying through mass queries, showing how blurry the line is between normal competitive learning and prohibited model extraction.

• Google is rolling Gemini into cars with Google built-in, starting in the U.S. with English-language support and expanding over the coming months. Lauren Forristal reports that GM says Gemini is coming to about 4 million 2022-and-newer vehicles across Cadillac, Chevrolet, Buick, and GMC, while compatible existing cars can receive the assistant through software updates.

CONTENT

How I Do Content Engineering With Claude Code

Ryan Law explains how Ahrefs rebuilt its content production process around Claude Code, 23 custom skill files, and a workflow that generates publish-ready drafts in six to twelve minutes. The system has produced around 15 new articles and updated about 30 more, but Ryan frames it as a way to maintain an evergreen content library rather than flood the blog with AI posts. The workflow mirrors Ahrefs' human editorial process by chaining skills for keyword research, gap analysis, outlining, drafting, troubleshooting, data sourcing, and final formatting.

The One Prompt That Improves Any Page by Checking for 5 Types of Specificity

Andy Crestodina explains how vague website pages fail visitors, search engines, and AI systems by missing the details that make an offer clear. His specificity audit checks five high-impact elements: the H1, hero image, subheads, navigation labels, and calls to action. The prompt scores each element from 1 to 5, flags generic words like "solutions," "services," "learn more," and "get started," then asks the AI to rewrite weak copy instead of giving loose advice. Andy also adds a CTA path check, asking the AI to follow the main call-to-action link and compare the button promise against the destination page.

SOCIAL MEDIA

I Got Shadowbanned on YouTube to Prove It's Not Luck

vidIQ tests YouTube's shadowban debate by creating a brand-new Shorts channel called It's AI Dog Life and filling it with prompt-generated dog videos. The first short reached about 2,500 views in 24 hours, but only 500 counted as engaged views, and the stayed-to-watch rate sat at 25%. Across the channel, stay-to-watch never climbed above 35%, sometimes dropped to 15%, while vidIQ says its main channel sits around 70% and top Shorts channels often exceed 80%. By the 26th video, uploads were getting single-digit views or zero views, which vidIQ ties to YouTube identifying the format as inauthentic content: simple prompt templates, little variation, and easily repeatable output at scale.

My Take: This is the clearest social platform lesson in the AI content boom: volume does not matter if the feed can tell viewers do not care. I would watch stayed-to-watch and engaged views before judging any Shorts experiment. If the sample audience keeps rejecting the format, the fix is not more posting. It is a different creative premise.

Reddit Marketing for SaaS: Insights From 117 Brands

Victory Umurhurhu analyzed 117 SaaS brands across 11 categories and more than 300 Reddit threads to see how the platform shapes brand discovery, sentiment, and AI search visibility. She found that 30 brands had no Reddit presence, another 23 had abandoned subreddits, and conversations still continued through user questions, comparisons, complaints, and alternative recommendations. The strongest brand participation came from moderators who answered directly, admitted limitations, and avoided corporate copy. Victory also frames Reddit as customer research, surfacing onboarding friction, integration issues, mobile complaints, AI feature frustrations, pricing concerns, workflows, screenshots, tutorials, and competitor gaps.

Social Ripples

• X rebuilt its ad platform around xAI models, calling it the biggest ad system refresh in the company's history. The update focuses on simplicity, advertiser control, and AI relevance matching, with X saying modern retrieval and ranking systems can improve targeting, relevance, engagement, and ROI.

• Instagram is expanding originality protections to photos and carousels, applying the same recommendation rules it introduced for Reels in 2024. Accounts that mainly repost unoriginal photos, carousels, or Reels will lose recommendation eligibility until most recent posts are considered original over a rolling 30-day period.

NEWS

Q1 2026 Earnings Call: Remarks From Our CEO

Sundar Pichai says Alphabet's Q1 2026 growth is being driven by AI across Search, Cloud, subscriptions, and models. Search & Other Advertising revenue grew 19%, queries reached an all-time high, and Google says AI Overviews are driving overall Search growth while AI Mode is seeing strong global growth in users and usage. Google also reduced Search latency by more than 35% over five years and cut the cost of core AI responses by more than 30% after moving AI Overviews and AI Mode to Gemini 3. Cloud revenue grew 63% to more than $20 billion, backlog nearly doubled quarter over quarter to over $460 billion, Gemini Enterprise paid monthly active users grew 40%, and first-party models now process more than 16 billion tokens per minute through direct API customer use.

My Take: These numbers show why AI search is not just a product experiment anymore. If you didn’t know already, Google is tying AI features directly to Search usage, ad growth, infrastructure efficiency, and enterprise adoption. For SEOs, that means AI visibility is moving into the same business system as classic search, not sitting off to the side as a separate channel.

Microsoft Bing Ads Revenue Up 12%

Barry Schwartz reports that Microsoft search advertising revenue excluding traffic acquisition costs grew 12% in Q3 2026, up 9% in constant currency. Bing also reached 1 billion monthly active users, a milestone Microsoft employees framed as a major signal for SEOs watching search behavior shift beyond Google. Microsoft reported $82.9 billion in quarterly revenue, up 18%, operating income of $38.4 billion, up 20%, and net income of $31.8 billion, up 20% on a non-GAAP basis. Satya Nadella also said Microsoft's AI business passed a $37 billion annual revenue run rate, up 123% year over year, tying the quarter's search growth to the broader agentic computing push.

My Take: If Microsoft keeps growing both usage and ad revenue, ignoring Bing becomes harder to justify for SEO and paid search teams.

WAYS WE CAN WORK TOGETHER

Floyi - Build Topical Authority that wins in Google and AI Search. 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

UGC That Drives Conversions — at Scale

minisocial helps top brands produce UGC that converts — top 1% CVR on TikTok, 50% lower cost per add-to-cart, 30%+ ROAS lift. Brands like Plant People, immi, Imperfect Foods, and Topicals are seeing results they couldn't replicate in-house. Create your brief in 10 minutes, approve your curated creators, and download scroll-stopping content. No long-term commitments, no hassle.

Get Started Now