In partnership with
Your weekly guide on everything AI marketing. Covering everything from AI marketing news, tips, deep dives, events, podcasts, jobs and much more. Never miss a beat with our 5-minute newsletter.
In today’s email:
Gartner finds half of consumers think GenAI made content worse
A German court makes Google liable for its AI answers
GSTV swaps demographics for agentic, behavioural targeting
Plus: why a content chief says stop treating AI like a factory, a 20-minute brand-voice workflow, and the week Washington switched off the most powerful Claude models.
The Top 3 Stories
1. Gartner: half of consumers say GenAI made content worse
A Gartner survey presented at its Marketing Symposium found 49% of US consumers think GenAI has worsened the quality of content they see, rising to 57% among Gen Z and millennials. The warning is that brands flooding channels with automated output risk severe marketing fatigue. Volume is not value.
Why it matters: AI is great for scale, but pairing it with a clear quality bar is what keeps audiences leaning in. 📰 CX Today
2. A German court makes Google liable for its AI answers
A Munich court ruled Google is directly liable for false statements in its AI Overviews, finding the summaries are Google's own words rather than a neutral list of links. The case began when an Overview tied two publishers to scams that did not exist. It is among the first rulings on who owns AI mistakes.
Why it matters: The ruling is an early sign that accuracy and oversight will matter more as AI-generated content becomes mainstream. 📰 Engadget
3. GSTV ditches demographics for agentic targeting
GSTV became the first media network outside Stagwell to adopt the Stagwell Agentic Targeting System, connecting its location footprint to an ID graph spanning 260 million-plus US consumers. Conagra is the first advertiser using it, moving from static demographic assumptions to real-time behavioural audiences. Agentic targeting goes mainstream.
Why it matters: Behavioural, real-time audiences beat age-and-gender guesses, and the brands testing them now set the benchmark. 📰 Marketing Dive
More trending AI marketing news from last week
Contentstack ships its Agentic Experience Platform for enterprises.
Koddi: 84% of commerce leaders will fund AI-shortlist visibility.
WPP Media forecasts AI search ad revenue will top $100bn by 2030.
BCG: 90% of CMOs say GenAI is reshaping discovery, but only 8% run autonomous agents.
Hearst debuts AURA IQ, agents draft RFP responses in minutes.
Your growth team woke up to a briefing they didn't ask for.
Monday 7am. Three messages in #growth.
Stripe revenue by channel, Meta and Google spend reconciled against GA4, Klaviyo flow performance, Shopify AOV by source. Posted by Viktor at 6am.
The campaign brief he wrote sits in #campaigns. Brand monitoring scrape runs every six hours. Competitor pricing update lands every Friday.
Your media buyer, content lead, and CMO open Slack to the same prepared room. 3,000+ integrations including every ad platform, CDP, and CMS you run.
"Viktor is like the most capable all-round colleague you can imagine." Sam, CEO, Givr.
Quote of the week
Stop treating AI like a content factory.”
Trending AI tools for marketers
Here are 4 trending tools to explore this week. Quick, practical upgrades to your AI marketing arsenal:
Octolane: Chat-first self-driving CRM that logs deals and drafts follow-ups for you.
Databox: Ask your live marketing metrics questions straight from Claude or ChatGPT.
mailX by mailwarm: Deliverability toolkit that tells you why emails hit spam.
Context.dev: One API gives AI agents clean brand and web data.
TV, podcasts & streaming
Marketing Over Coffee: UChicago Medicine CMO Andrew Chang at Salesforce Connections
Hosts John Wall and Christopher Penn sit down with a hospital CMO on building a modern data foundation, going all-in on Data 360, and where AI actually fits. The sharpest thread: trustworthy data, not clever prompts, is what makes AI usable in a high-stakes category. For marketers, it is a useful reminder that AI output is only ever as credible as the first-party data feeding it - listen to it here.
AI training
How to brief a self-driving CRM using Octolane
The overview:
Octolane turns plain-language requests into logged deals, follow-up drafts and pipeline updates, so you spend minutes on admin rather than an afternoon.
Step-by-step:
Connect your inbox and calendar
Link Gmail and your calendar first. Octolane auto-detects deals from real conversations, so connecting both gives it the context to work from.
Ask in plain English
Type a request like "show me deals that haven't moved in 10 days." Use natural phrasing, not filters; the tool is built to parse intent, not query syntax.
Draft a follow-up
Ask it to "follow up with David about last week's pricing question." Check the draft before sending, since AI tone can run slightly generic on first pass.
Let it log the call
Use the meeting recorder so recaps write straight into the deal. Good output looks like a clean summary with next steps, not a raw transcript.
Review the pipeline
Drag deals across the kanban view to sanity-check what the agent inferred. Catching a misfiled deal early keeps your reporting honest.
Pro tip: Connect Octolane's MCP server to a tool you already use so the same context follows you across apps.
The weekly deep dive
Washington just switched off Anthropic's most powerful Claude models
For the first time, the US government forced a frontier AI model offline, and it happened three days after launch.
What happened
Anthropic released Claude Fable 5 and Mythos 5 on the 9th of June, then suspended both 3 days later after an emergency export-control directive barring use by foreign nationals. Unable to screen non-US users in real time, Anthropic pulled the models for everyone. It is the first government-forced takedown of a publicly deployed frontier model.
Why it happened
The directive cited a method for jailbreaking Fable's safeguards to extract information useful for cyberattacks. Anthropic disputes the severity, noting the same capability sits in other public models including GPT-5.5, and warns that applying this standard industry-wide would stall every frontier launch.
What happens next
There is no restoration timeline, and Anthropic has not said whether resolution needs an independent review of the flaw, a patch, or a slower legal fight over the directive itself. Analysts expect the precedent to spread: pre-launch government security reviews, licensing for advanced models, and capability tiers rather than universal access. The risk is no longer specific to one vendor; any model that crosses a capability threshold could be next, which makes model availability a planning assumption rather than a given.
Our takeaway: Treat model access as infrastructure that can fail, keep a tested fallback wired in, and watch for pre-launch review rules that could delay the tools you are planning around.
📖 Read the full article at VentureBeat
