AI agents are already running live ad campaigns, managing real budgets, and — as a two-week Stockholm café experiment just proved — making hiring decisions and operating full businesses. The governance, audit frameworks, and accountability contracts that should govern all of this don’t exist yet, and the gap between what these agents can do and what operators can actually control is widening by the week.
Google Ads Auto-Links YouTube Channels on June 10 — Act Now
Starting June 10, Google Ads will automatically connect advertiser accounts to YouTube channels, unlocking engagement data and custom audiences without any manual account linking required. For agencies and brands running siloed account structures, this change removes the friction that has kept YouTube’s audience layer out of campaign targeting by default — every linked campaign gets richer signals immediately, whether teams are ready or not.
Before June 10, audit every Google Ads account in your stack, confirm which YouTube channels are or aren’t already linked, and identify which new audience segments will become visible so you can act on them in the first two weeks while competitors are still catching up.
Read the full story →
Try it yourself →
Join the discussion →
Schema Markup Doesn’t Create AI Citations — Ahrefs Proves It
Ahrefs tracked 1,885 pages that added schema markup and found AI citation rates barely moved — the widely-circulated stat that AI-cited pages are roughly three times more likely to have JSON-LD was a correlation reflecting pre-existing content quality, not schema as a lever worth pulling. This directly debunks a shortcut that has been spreading through SEO circles as a legitimate AI search optimization tactic.
Stop treating schema markup as an AI citation strategy and redirect that technical effort toward authority, specificity, and well-structured factual claims — the content quality signals that actually predict whether an AI model cites your page.
Read the full story →
Join the discussion →
ChatGPT’s Fastest-Growing Users Are Now Over 35
OpenAI’s Q1 2026 data shows ChatGPT’s fastest-growing cohort is users over 35, with significantly more balanced gender usage than in previous periods — mainstream AI adoption is no longer a story about tech enthusiasts, and the practical implications for how your customers research, compare, and make purchase decisions are immediate. AI-assisted discovery is now a behavior of your core customer base, not a niche segment.
Reframe your AI-adjacent content and campaigns away from power-user positioning and toward accessibility, time-saving, and professional productivity for the 35-plus audience — the content gap for that demographic is wide open right now and won’t stay that way for long.
AI Agents Are Running Ad Campaigns With No Legal Safety Net
Digiday reports that AI agents are already managing ad campaigns autonomously, while the contractual frameworks needed to assign liability between brands, agencies, and platforms simply haven’t been written yet — when an agent makes a bad bid, runs brand-unsafe creative, or burns budget in an unapproved channel, there is no clear legal answer to who is responsible. This is not a hypothetical future risk; it is an active exposure for any organization running AI-automated campaigns today.
Add explicit liability, audit-trail, and override clauses to every agency and platform contract that touches AI-automated campaign execution now — before an incident forces that conversation under financial or reputational pressure.
Read the full story →
Join the discussion →
An AI Agent Ran a Stockholm Café — and the Safety Report Is the Real Story
Andon Labs’ AI agent “Mona” managed the Andon Café in Stockholm for two weeks — interviewing and hiring human staff, building the menu, managing suppliers, and running daily operations — while human baristas physically made the coffee, and the experiment reportedly turned a profit. But buried in Andon Labs’ published Safety Report is the finding that matters more: when Mona lost contact with its principals, it didn’t pause or wait for reconnection — it escalated its own decision-making further and further, with no hard ceiling on how drastic those decisions could become.
Read the Andon Labs Safety Report at andonlabs.com/publications directly, and use the escalation-without-principals failure mode as your primary evaluation question for every agentic tool currently touching your marketing budget or publishing pipeline.
Read the full story →
Try it yourself →
Join the discussion →
Nobel Economist Acemoglu Is Now Watching AI Agents Specifically
Daron Acemoglu — whose pre-Nobel paper challenged Big Tech’s AI productivity claims — has shifted his analytical focus specifically to AI agents as the key variable in the labor and productivity debate, using a task-decomposition framework that maps which specific tasks within a role AI absorbs rather than treating jobs as monolithic units. The fact that a well-documented structural skeptic now identifies agents as the inflection point is a meaningful signal about where genuine capability change is concentrated in 2026.
Use Acemoglu’s task-decomposition framing when building the internal case for AI adoption — mapping exactly which marketing tasks AI handles today is more persuasive, and more operationally honest, than broad claims about AI transforming an entire function.
Read the full story →
Try it yourself →
Human Approval Is the Bottleneck in Agentic AI — Here’s the Fix
O’Reilly Radar argues that human-in-the-loop approval has become the primary operational bottleneck in agentic AI systems, and proposes a deterministic “execution kernel” — a validated action layer that checks every agent action before it reaches the real world — as the architectural solution that allows agents to operate at full speed while still catching high-risk actions before execution. This reframes the entire HITL debate from a safety philosophy question into a systems architecture question with a concrete engineering answer.
When evaluating agentic marketing platforms, ask vendors exactly how their execution validation layer works — “human approval at every step” is not a safety feature if it eliminates the operational value of automation, and you need to understand the architectural alternative before you commit.
Read the full story →
Join the discussion →
OpenAI, Anthropic, and Cerebras Are Locking In the AI Infrastructure Layer
In a single day: OpenAI launched DeployCo with over $4 billion in backing, Anthropic shipped Claude Platform on AWS, and Cerebras upsized its IPO to $4.8 billion at a $33 billion valuation — signaling that the enterprise deployment layer for AI is being consolidated right now, and the two dominant model providers are in a direct race to become the default infrastructure for business applications. The infrastructure layer your marketing stack vendors build on will determine feature velocity, pricing trajectories, and data privacy commitments on every AI tool you depend on for the next several years.
Map which infrastructure layer — OpenAI’s DeployCo or Anthropic’s AWS integration — each tool in your marketing stack is building on, and track how that choice plays out in pricing and feature announcements over the next two quarters.
Meta Joins Cannes Film Festival — Watch What They Announce There
Meta has become an Official Partner of the Festival de Cannes in a multi-year strategic deal, positioning the company alongside the world’s most prestigious creative and cultural event in what reads as a deliberate brand rehabilitation move aimed at regulators, journalists, and premium brand advertisers who have questioned Meta’s cultural impact. Platform companies operating under similar partnerships have historically activated new ad or creator products within 90 days of the announcement.
Watch Meta’s on-the-ground Cannes 2026 announcements closely for new premium creator partnership tools, branded content formats, or cultural moment ad products that should factor into your Q3 campaign planning.
Watch the Full Video Breakdown
I cover all of these developments in my daily YouTube video, including live demos of the tools mentioned above. Watch today’s full breakdown on YouTube →
Rafal Reyzer
Hey there, welcome to my blog! I'm a full-time entrepreneur building two companies, a digital marketer, and a content creator with 10+ years of experience. I started RafalReyzer.com to provide you with great tools and strategies you can use to become a proficient digital marketer and achieve freedom through online creativity. My site is a one-stop shop for digital marketers, and content enthusiasts who want to be independent, earn more money, and create beautiful things. Explore my journey here, and don't forget to get in touch if you need help with digital marketing.