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:
L'Oréal wires its own content engine and try-on tools into generative AI
Klarna's CMO builds an AI clone of himself for staff to vent at
A price war hits the AI industry before any bubble can burst
Plus: a former Mastercard CMO calls this a "golden era," previews of the Cannes conversations that actually matter, and Lastly… AI Marketers is now officially…LIVE!
The Top 3 Stories
1. L'Oréal feeds its own data into the machines
At VivaTech, L'Oréal detailed CreAItech, an in-house generative platform that produces brand-true imagery and video, plus plans to bring Maybelline's ModiFace virtual try-on into conversational AI. It will feed verified product data straight into the models behind major assistants. The aim is discovery grounded in first-party data, not scraped reviews.
Why it matters: Owning the data an AI cites about your brand is fast becoming a core marketing job, not an IT one. 📰 Glossy
2. Klarna's CMO builds an AI clone to absorb the complaints
Klarna CMO David Sandström created an internal AI replica of himself, trained to sound like him and stay endlessly apologetic, so colleagues could vent during budget cuts instead of sending angry Slack messages. He described it on an ElevenLabs webinar, saying he "just didn't want to hear the whining in the meetings anymore." Klarna later built a separate chatbot modelled on its CEO for customer feedback.
Why it matters: If a CMO will clone himself to skip the complaints, expect customer-facing AI personas next, and the trust questions that come with them. 📰 Entrepreneur
3. A price war hits AI before the bubble can burst
Forbes argues a price war is reshaping AI economics well ahead of any bubble bursting, as enterprises slam the brakes on runaway token spending. Anthropic cut Claude Opus pricing 67% at its 4.5 launch and OpenAI added a 50% Flex discount, with more cuts expected. One example: Uber burned through its 2026 AI coding budget in four months.
Why it matters: Falling token prices mean the AI tooling in your stack is about to get cheaper, so lock in flexibility rather than long contracts. 📰 Forbes
More trending AI marketing news from last week
Warner Bros. Discovery rebuilt its ad-tech stack around agentic AI and AWS.
Pinterest launched a standalone AI shopping app, Ask Pinterest, and opened its ads to agents.
Snap rebuilt its ad stack with an AI Smart Assistant and MCP server.
Microsoft detailed new AI ad tools at Activate 2026.
Pew found 49% of US adults now use AI chatbots, but trust is lagging.
AI Marketers is now officially… LIVE!
The newsletter you've been reading is now the home of AI-native marketing: a full platform, free to use, built for marketers building with AI.
What's now live:
A free AI Audit. one short survey and you get back an AI maturity score, AI gaps, tailored playbooks and a 90-day roadmap.
An AI tool directory. Look up vetted tools by what you actually need them to do.
A jobs board. AI marketing roles, in one place.
Events. The conferences, webinars and meetups worth your time, online and in person.
The weekly newsletter. Still here, still landing every week.
And every signup gets a portal: save tools to your stack, jobs you like the look of, and events you want reminders for, all in one place.
Because you read the newsletter, your account is already reserved. One click, same email, no password.
Free, always. You'll keep getting the newsletter either way.
Quote of the week
This is the golden era that we are about to enter as far as marketing is concerned.”
Trending AI tools for marketers
Here are 4 trending tools to explore this week. Quick, practical upgrades to your AI marketing arsenal:
Honestly: Surfaces and verifies real customer conversations across Reddit, TikTok and X.
Brieflee: Spell-check for video that catches ad quality issues before you post.
Gauge: A marketing agent unifying organic, paid and AI-search data.
HookLens: Diagnoses your video hooks and flags fixes before launch.
TV, podcasts & streaming
Cannes 2026: what everyone is talking about
eMarketer's analysts preview the Cannes sessions worth caring about, and the most revealing thread is what AI is quietly doing to creative teams. As automated campaign tools demand ever more assets to feed them, one brand in a Meta case study went from three pieces of content a week to over fifty, and the supply gap creative teams have lived with for a decade has widened into a chasm. The uncomfortable conclusion for marketers: when AI makes volume effectively free, the scarce thing becomes an idea worth watching - watch it here.
AI training
How to find honest customer feedback using Honestly
The overview:
Honestly pulls verified conversations about your brand from across social platforms and turns them into themes you can act on, in well under 30 minutes.
Step-by-step:
Add your product and competitors
Enter your brand plus two or three rivals.
Naming competitors lets you benchmark sentiment rather than read your own mentions in a vacuum.Set your platforms and window
Choose the networks where your buyers actually talk and a 30 to 90 day window.
A tighter window surfaces reactions to recent launches; a wider one shows durable themes.Run the scan and filter the noise
Let the tool verify posts and strip out bots and sponsored content.
Skipping this step is the usual mistake; raw mentions are mostly spam and astroturf.Cluster the themes
Group results into recurring praise, complaints and feature requests.
Good output reads like a backlog, not a word cloud.Pull three quotes per theme
Lift real customer language for ad copy and landing pages.
Their words convert better than yours, so keep them verbatim.
Pro tip: Re-run the same scan monthly and track which complaints shrink after you ship, so feedback feeds a loop rather than a one-off report.
The weekly deep dive
Synthetic audiences just crossed a line marketers said they never would
For years, media buyers used AI-built "synthetic" audiences for planning but refused to spend real programmatic money against them. That line is now moving.
What WPP Media actually did
WPP Media partnered with a firm called Givsly to build audience segments from an unusual signal: charity-donation patterns by postcode, blended with census and behavioural data. Using a Claude-based system, it turns a brief like "Californian women interested in wellness" into a values-led segment, then layers in WPP's own first- and second-party data. Crucially, those synthetic profiles were used in live private-marketplace deals, not just slides.
Why the resistance was rational
Buyers held the line because a modelled audience is a guess until a real human converts, and pointing programmatic budget at a guess risks paying for impressions nobody wanted. What is shifting their stance is early evidence: across four beauty and fashion campaigns on CTV and online video, the approach lifted video completion rates by 2%. Small, but enough to move a pilot toward a rollout.
The signal for marketers
Targeting by stated values, rather than just demographics or past behaviour, is a genuinely new axis, and one consumers may find either flattering or unsettling. The 2% lift is the headline, but the real story is precedent: once one holding company spends real money against synthetic audiences and shows a result, the rest will be asked why they are not.
Our takeaway: Before synthetic audiences reach your plan, decide where you stand on targeting people by their values, because the capability is arriving faster than the ethics debate around it.
📖 Read the full article at Digiday
