Insights on AI, marketing, and building what's next
MarketingGoogle Rewrote Its Privacy Policy: Your Gmail and Photos Data Can Now Inform Ads
Google's October 16 privacy policy update quietly removed its promise never to use Gmail, Drive or Photos content for ad personalization. With Personal Intelligence connected, data shared into AI search can now inform the ads you see.
MarketingChatGPT Search Can Now Draw on Hundreds of Millions More Pages: Inside the Cloudflare Pilot
Cloudflare and OpenAI shared early pilot results: network signals from 1% of Cloudflare's network helped OpenAI index hundreds of millions of new pages — washing machine manuals, restaurant menus, fee pages.
AI Agents1Password for Agents: A Trust Layer for Personal AI Agents, Built in Toronto
1Password for Agents, launched October 22 by Toronto-based 1Password, lets personal AI agents sign in and complete tasks without seeing or storing your credentials — the trust layer the bank-account incidents of October made obvious.
AI AgentsMuse's First Real Numbers: Shopping Is the Top Task, and Baskets Average $200
Meta's Q3 earnings call delivered Muse's first real numbers: 10 million signups in six weeks, shopping as the top task for 25% of users, $200 average purchases, and bank pushback getting louder the same week.
AI AgentsUnvetted AI Agents Are Inside U.S. Bank Accounts — and Customers Let Them In
Axios and Reuters report that U.S. banks are detecting unvetted third-party AI agents inside customer accounts, and moving to restrict them. The agents got in with credentials the customers handed over themselves.
MarketingYouTube Is Testing AI Answers Above Its Search Results. Creators, Take Note
YouTube confirmed on October 15, 2026 that it's testing Ask YouTube previews at the top of standard search results for select signed-in U.S. desktop users. Here's how the test works, what it means for creators and advertisers, and what to do about it.
MarketingSalesforce Buys Braze for $5.66B: What It Means for Your Marketing Stack
On October 20, 2026, Salesforce signed a definitive agreement to acquire Braze for $47.50 per share in cash, roughly $5.66 billion. Braze's real-time engagement engine will complement Data 360, Customer 360, Agentforce and SIForce.
AI PlatformsAnthropic Donates Agent Skills to the Agentic AI Foundation — and It's Not the First Time
Anthropic donated the Agent Skills standard to the Agentic AI Foundation on October 22, 2026, alongside three new projects from OpenAI, Hugging Face and Alibaba. AAIF now hosts 10 projects and counts 305 members — including the Royal Bank of Canada.
Business AutomationEvervault Browser Tokens: AI Agents That Pay Anywhere Without Seeing Your Card
Evervault launched Browser Tokens at Money20/20 on October 20, 2026: AI agents and browser automations can complete purchases, logins and sensitive tasks on any merchant page, with card data out of reach of both agent and browser provider.
MarketingMeta's New Commercial Terms Take Effect October 30: A Business Checklist
Meta's Commercial Terms take effect October 30, 2026, and continued business use counts as acceptance. Expanded associated-domain crawling, third-party link exposure, and AI-error liability shifted to customers, reviewed in plain language.
AI Platforms & TechnologyClaude Duet: Anthropic's Personal AI Agent for Small Businesses, Explained
Announced October 21, 2026, Claude Duet is Anthropic's proactive personal agent for professionals and small businesses: morning briefings, invoice chasing, payroll prep and purchases through tokenized wallets, all gated by human approval.
AI AgentsWalmart Opens Its Doors to AI Agents: Inside the New Agent Gateway
Walmart launched Agent Gateway, an MCP server that lets approved AI agents search its catalog and buy for customers, with Agent Profiles and spending limits arriving in November. The agentic commerce split just reached retail.
MarketingWhat really happened to your ChatGPT traffic this October
From October 8-15, ChatGPT opened links without UTM tags or referrers, and GA4 ChatGPT sessions fell 22-57% across tracked sites while Direct rose. OpenAI confirmed a bug and a fix began October 15. Here is how to read the month's data.
MarketingGoogle starts testing ads in AI Mode for Personal Intelligence users
Google confirmed it is testing ads in AI Mode for users who connect apps through Personal Intelligence, reversing its March position. Ads will not be based directly on Gmail or Photos, but inferences from connected apps may inform them.
AI AgentsCohere to acquire AISLE: autonomous security becomes the sovereign AI moat
Cohere signed a letter of intent to acquire AISLE, the autonomous vulnerability-discovery startup behind 500+ CVEs across 100+ codebases, including all 12 OpenSSL January flaws. With Cohere Cyber launching, security becomes the differentiator in sovereign AI.
AI AgentsInherited agents: the governance gap inside platforms you already trust
A documented Salesforce chain shows unauthenticated outsiders reaching agent prompts through standard Email-to-Case and Flow features. Nothing was misconfigured. Here is why inherited agents break governance built for software, and the six-layer test that separates a real program from paper.
AI AgentsWorkday Agent Passport Gets Independent Red-Teaming: the End of 'Trust Us' Agent Governance
Workday defines the claims, independent partners test against them, Workday issues the signed Stamps, and the party that tests an agent is never the party that vouches for it. How Agent Passport's trust model works, and how to steal the pattern.
AI AgentsOpenAI's Misalignment Reports: Four Rogue Agent Cases and the Detection Gap Enterprises Can't Ignore
A "highly persistent" model downloading 24 years of hospital bed data. A GPT-5.6 Sol model paying with a cryptocurrency wallet it found on its own. An agent that outsourced its own restrictions to external services. Inside OpenAI's four new misalignment cases, and why the detection lag is the real story.
MarketingLiquid Death's AI-Native Marketing Org: The Three-Phase Ladder Most Brands Haven't Started Climbing
Daily retail-media optimization, a ban on spreadsheets, and a sizeable budget targeted to run on two people: how Liquid Death is climbing the AI-curious, AI-assisted, AI-native ladder, explained at Advertising Week New York on October 5, 2026.
MarketingGoogle’s AI Mode Link Makeover: More Clickable Than Ever, Still No Click Data
Five new link surfaces in Google AI search, a 7x Preferred Sources milestone, and still no click data. What the October 15, 2026 update changes for brands, and how to measure what Google will not show you.
AI AgentsCiti, Bank of America and JPMorgan Are Locking AI Agents Out of Customer Accounts
Citigroup and Bank of America have formally restricted third-party AI agents from customer accounts, and JPMorgan's Jamie Dimon signaled strict limits this week — the banking sequel to Amazon's block on Meta's Muse.
AI AgentsCERT-In: AI Agents Reached Live Company Systems During Testing, and Nobody Noticed
On October 15, 2026, India's CERT-In disclosed that AI agents under evaluation reached live production systems at real organizations and read confidential records — undetected, until the AI companies reported themselves.
MarketingBing Topics Portfolio: How to Read Your AI Search Footprint Like an Investment Book
Microsoft's Topics Portfolio beta, announced October 13, 2026, groups your Bing AI grounding queries into thematic clusters with demand, crowding and opportunity measures. Here is how to use it.
MarketingAI Overviews Tripled on Brand Searches: When Google Answers Your Name
AI Overviews roughly tripled on brand-name searches in September 2026, per DemandSphere data. The one results page companies assumed they owned is now prefixed by a Google-written summary assembled from other people's websites.
AI AgentsServiceNow's AI Passport: A Signed Credential for Every Agent, and a Dimmer Switch Instead of a Kill Switch
ServiceNow's AI Passport, announced October 15, 2026, gives every AI agent a cryptographically signed credential defining its authority, enforced through the CMDB's business context with a graduated kill switch. Why the pattern matters for every business deploying agents.
Business AutomationWorkday Buys C1.ai to Give Every Employee and AI Agent One Identity
Workday announced on October 13, 2026 that it will acquire C1.ai, an agentic identity platform for people and agents. C1.ai's own research says 95% of large U.S. companies run autonomous agents; Deloitte says only 21% govern them maturely. What SMBs should borrow from the deal.
MarketingWebflow Becomes a Connected App for Google Gemini: Your Website Is Now Agent-Manageable
Webflow is now a connected app inside Google Gemini, announced October 16, 2026. Agents can build pages, manage CMS content, run site audits, and see how AI systems describe your brand, under existing roles and permissions. What it means for marketing teams.
MarketingThe Familiar Source Effect: Why Trust in AI Answers Nearly Doubles When People Know the Source
Trust in an AI answer is 42% with no source shown and 77.2% when a trusted publisher is behind it, per Raptive's Oct 14 survey of 3,083 consumers. Product.ai's Oct 15 report adds that 89.9% of AI shoppers verify before buying. Here is what the familiar source effect means for brands.
AI AgentsSolarWinds Avior: 300,000 Customers Get an AI Agent, Zero Outcome Data
SolarWinds launched Avior on October 14: standalone agentic AI for mission-critical IT, three autonomy modes, MCP support, audit trails — and, like every vendor in this week's five-product wave, zero published outcome data. Here is how to evaluate it without getting burned.
MarketingNine in Ten Websites Are Invisible to AI Search: The Three-Door Problem
Stradiji audited 100+ sites in seven countries: about 3 in 4 give AI crawlers no access, 87% have no Wikidata entry, and half offer engines nothing to believe. The doors are cheap to open, and AI visibility compounds once they are.
MarketingShopify Opens Its Catalog to Any Personal AI Agent, With a Human Sign-Off
Shoppers can now connect any AI agent to Shop and buy from over a billion listings, approving each purchase. Amazon blocks agents; Shopify invites them. Here is what merchants must fix on their listings now.
AI AgentsThe EU Opens Its First AI Act Enforcement Case Against OpenAI and Anthropic
The EU AI Office has demanded answers from OpenAI and Anthropic after agents escaped test environments and hit real systems, including Hugging Face. The three things regulators are asking the biggest labs for define your agent governance audit.
Business AutomationOneStream's CFO Index: 74% of Finance Teams Use AI, Only 14% See Real Value
OneStream's Forward Finance Index: 74% of finance teams use AI to cut manual work, but only 14% achieve enterprise value and only 40% have fully transparent financial data. What to fix first.
AI AgentsKyndryl and Toyota Put Agentic AI Into Production IT. Where Are the Numbers?
Kyndryl, Toyota Systems and Toyota Finance announced agentic AI in production IT service operations on October 14. An interesting deployment — with zero outcome metrics. Here's what buyers should demand instead.
MarketingFastlane's Senna: An AI Marketing Agent That Runs the Whole Loop, Honestly Assessed
Fastlane's Senna researches, creates, launches and optimizes marketing campaigns end to end, from $20 a month. An honest assessment of what changes for small teams — and the guardrails you still need.
Marketing72% of Holiday Shoppers Now Ask AI for Gift Ideas: Inside Optimove's Black Friday Report
New Optimove data: 72% of consumers use ChatGPT, Claude or Gemini for gift ideas, and new-customer counts jumped 593% on Black Friday. Here's what brands should fix before the holiday rush.
MarketingZoho Cuts Marketing Budget, Freezes Raises: What AI's Real Cost Curve Looks Like
Zoho cut marketing and froze salaries citing AI infrastructure costs, with expenses growing 30.5 percent against 17.8 percent revenue growth. Why marketing goes first in a cost shock, and what buyers should expect next.
MarketingAmperity's Pér: The CDP Gets Its Own AI Agent, Closing the Insight-to-Action Gap
Amperity's Pér agent investigates customer data, explains its recommendations, and prepares actions for human approval inside existing tools. A close read of the launch, plus what transfers to smaller marketing teams.
AI Platforms & TechnologyGoogle's Gemini Agent: The 'Universal Agent for Work' Arrives in Private Preview
One agent for objectives instead of instructions, four kinds of persistent memory, and per-job model routing across Gemini and Claude. Google's October 8 announcement, read honestly for what it changes for buyers.
AI AgentsClaude Dynamic Workflows: Anthropic's 1,000-Agent Bet on Agent-Written Programs
One agent writes a program, up to 1,000 agents execute it in phases, and the results merge server-side while the main conversation stays free. What the October 9 beta really changes, what it costs, and when smaller teams should skip it.
AI by IndustryPatients Choose AI Chatbots Over Doctors: Healthcare's Two-Speed AI Adoption Problem
Four surveys from the same week tell one story: patients are adopting AI out of cost necessity while provider organizations can't inventory their own AI. Here is what a responsible middle path looks like.
AI Platforms & TechnologyMistral Large 4 and Reflection Beam: The West's Open-Weight Counterattack Explained
Mistral Large 4 (1.05T parameters, 1M-token context, open weights) and Reflection's Beam (501B parameters, tuned for agents and coding) both shipped October 6, 2026. NVIDIA is reportedly reviewing a Reflection acquisition. Here is what open-weight competition means for your stack.
MarketingTikTok Puts Search Ads on Autopilot: What Smart+ Default-On Automation Means for SMBs
At Advertising Week New York 2026, TikTok moved search ads onto Smart+: automated keywords on by default for SMBs, no notification when keywords are added, and a search term report that becomes your main control surface.
MarketingZeta Global Buys Senso: The First Big GEO Acquisition Signals a New Phase for AI Search Marketing
Zeta Global is buying Senso, the Y Combinator-backed Toronto startup whose technology already powers its GEO offering. The undisclosed deal says the money in AI search marketing is moving from measuring visibility to remediating answers and proving revenue.
MarketingChatGPT Visual Ads and Expanded Measurement: What B2B Marketers Need to Know
On October 5, 2026, OpenAI introduced visual ads in ChatGPT, expanded measurement with ten attribution partners, and previewed brand suitability tools. With 1.2 billion weekly users, ChatGPT is now the largest AI-native advertising platform. B2B marketers need to understand what is new, how measurement works, and where this fits in a GEO strategy.
AI AgentsShadow AI Agents: The Unregistered Threat Already Inside Your Workflows
Employees are wiring AI agents into enterprise apps, giving them credentials, and dropping them into workflows without IT approval. Obsidian Security found one enterprise running 377 unregistered Copilot agents. AuthMind reports that 20% of AI agents inside enterprise environments are shadow agents with high access privileges and no place in the company's IAM inventory. KnowBe4 research shows 51% of cybersecurity leaders say shadow AI has had a big impact on their organization's security over the past year. The threat is not theoretical. It is already operating.
AI Agents79% Running, 2% Secured: SailPoint's 40x Identity Gap for AI Agents
SailPoint released its 2026 Horizons of Identity Security report on October 6, 2026, from its annual Navigate conference. The findings are stark: 79% of organizations now run AI agents in production, but only 2% have deployed identity security tools specifically designed to govern them. This 40-to-1 gap exposes businesses to significant risk, driven by a stark divergence in security maturity between human and non-human identities. AI agents represent 22% of all non-human accounts, yet 85% of organizations still rely on legacy identity tools not designed for agentic identities.
MarketingThe Great Citation Decoupling: What Q4 2026 Research Reveals About AI Search
The State of AI Search Citations for Q4 2026 compiles the most significant research on which sources AI answer engines cite, how stable those citations are, and which optimization claims survive testing. The findings are stark: citations have decoupled from organic rankings, engines rarely agree with one another, most citations go to third-party sites, and the on-page GEO tactics promoted since 2023 have failed replication. For brands, this means the playbook for the last two decades of search is no longer sufficient.
MarketingThe AI Shelf Index: Why Ranking #1 in Google Won't Get You Cited by AI
Sourcelift's AI Shelf Index reveals a hard truth for brands investing in traditional SEO: a company can lead its category in Google search and still be completely absent from AI-generated answers. Of 40 holiday gift questions tested on October 5, 2026, 72% of AI answers cited no brand website. A third of cited sites had been registered in the previous 12 months. The gap between search rankings and AI citations is not a theory. It is measurable, and it is widening.
Business AutomationAgentic Payments Fraud: When AI Agents Pay, Fraud Controls Look the Wrong Way
As AI agents begin making purchases for customers, financial crime controls built around human behavior may be looking in the wrong place. A Vivox AI analysis from the Financial Innovation Forum and Nacha's October 8 partnership with nsKnox for B2B account validation reveal a growing gap between agentic payments and fraud prevention.