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:
ChatGPT opens its first paid ad surface to brands, at a $60 CPM that prices out most of the room
Kana raises $15M and bets that flexible, human-supervised AI agents are what marketing teams actually want
Google Pomelli launches free AI product photography that reads your website and generates studio-quality shots
Plus: Sadoun's blunt verdict on AI ROI, this week's podcast on whether AI intensifies workloads, build a video ad from one sentence using Reloop, and four tools
The top 3 stories
1. ChatGPT Now Carries Ads and the First Placements Are Already Live
OpenAI began testing advertising in ChatGPT on 9 February, targeting logged-in adult users on its Free and Go tiers in the US. First-wave brand partners include Target, Ford, Adobe and Expedia, with placements appearing at the end of AI responses and clearly labelled as sponsored. The minimum advertiser commitment sits at $200,000 with a $60 CPM, roughly three times what Meta charges, placing this firmly in the enterprise bracket for now.
Why it matters: A new high-intent conversational ad surface has arrived. Users encountering placements are mid-conversation with an AI, not scrolling a feed, which changes how creative needs to be built and evaluated. 📰 TechCrunch
2. Kana Raises $15M for Flexible AI Marketing Agents
San Francisco-based Kana emerged from stealth on 18 February with a $15M seed round led by Mayfield. The company was founded by Tom Chavez and Vivek Vaidya, the team behind Rapt (acquired by Microsoft) and Krux (acquired by Salesforce), and deploys loosely coupled AI agents across audience targeting, campaign management, media planning and answer engine optimisation, integrating on the fly with existing marketing systems. A human-in-the-loop design keeps marketers reviewing and approving at each stage rather than ceding control to autonomous software.
Why it matters: Two martech acquisitions give these founders credibility most AI marketing startups cannot match. The emphasis on human oversight signals the market is not ready for agents that act without approval, and Kana is building accordingly. 📰 TechCrunch
3. Google Pomelli Launches Photoshoot: Free AI Product Photography for Brands
Google Labs rolled out a new feature for Pomelli on 19 February. Photoshoot uses the platform's Business DNA system to transform a single product image into polished studio-quality shots, including floating, ingredient, in-use and lifestyle styles, without any photography equipment or brief. Users can link directly to a product URL rather than uploading anything, and the feature remains completely free during the current beta.
Why it matters: Professional e-commerce photography costs anywhere from £400 to £4,000 per product. A free alternative that reads your website and generates variants is a direct challenge to that workflow, and a real opening for smaller brands. 📰 Google blog
See every move your competitors make.
Get unlimited access to the world’s top-performing Facebook ads — and the data behind them. Gethookd gives you a library of 38+ million winning ads so you can reverse-engineer what’s working right now. Instantly see your competitors’ best creatives, hooks, and offers in one place.
Spend less time guessing and more time scaling.
Start your 14-day free trial and start creating ads that actually convert.
More trending AI marketing news from last week
Digiday+ research found that 54% of marketing organisations have yet to deploy agentic AI in any workflow, with hallucination risk the chief structural blocker.
AI referrals to websites are up 357% year-on-year as answer engine optimisation becomes a required discipline alongside SEO.
Media agencies under WPP are testing AI planning agents but will not let them execute actual buys.
Digiday+ research found that 54% of marketing organisations have yet to deploy agentic AI in any workflow, with hallucination risk the chief structural blocker.
Perplexity dropped all advertising over trust concerns as rival AI platforms race to test it, citing research that 63% of users trust AI results less when ads are present.
Quote of the week
No doubt, AI is going to revolutionize everything we do but so far it is difficult to scale, expensive to put in place, and fails to deliver measurable value in 95% of cases."
Trending AI tools for marketers
Here are 4 trending tools to explore this week. Quick, practical upgrades to your AI marketing arsenal:
🎬 Reloop – Talk to an AI agent, describe your product, and get a polished video ad with an AI avatar, platform-optimised script and A/B-ready variants, with no camera, casting or editor required.
🎨 DesignLumo – Turns a text prompt into a fully editable, layered ad design with brand kit, typography and colour structure intact, skipping the template-hunting and rebuilding stage entirely.
✂️ CyberCut AI – Auto-slices long video footage into platform-ready social clips, generates precision captions and converts scripts into structured marketing videos in three steps.
🎯 Audience Loop – Builds and validates target audience segments before campaign launch, surfacing signals that inform creative and media decisions upfront rather than after spend.
TV, podcasts & streaming
Claude Cowork clearly explained (& how to use It for beginners)
Eliot Prince walks through Claude Cowork from scratch, covering installation, folder access, autonomous file tasks, invoice processing and browser integration, presented for anyone who has found recent AI releases genuinely overwhelming. A clean, step-by-step entry point for marketers who want to understand what agentic AI looks like in a real workflow, without the hype - watch it here.
AI training
Turn a product description into a polished video ad using Reloop
The overview: Reloop's AI agent handles scripting, avatar selection and format configuration from a simple conversational brief. The output is a branded video ad ready to export directly to Meta, TikTok or Google Ads.
Step-by-step:
Start your project.
Open Reloop and start a new project. In the chat interface, describe your product in one or two sentences, including who it is for and the single thing you most want viewers to remember. The more specific you are here, the better the script.Refine the hook.
Review the agent's suggested script and opening hook. If the hook does not feel right for your brand tone, ask for two or three alternatives in the same message. Reloop treats this as a single instruction, not multiple edits.Choose your avatar.
Select your avatar and setting. Choose a look and background that matches your target audience's context. Preview each option before committing, as the environment affects how the ad reads emotionally, not just visually.Set your format.
Choose vertical (TikTok, Reels), square (Meta feed) or horizontal (YouTube pre-roll). Reloop reformats pacing and layout automatically per platform.Request batch variants.
Ask Reloop to produce three versions with different opening five seconds. This gives you A/B test material without any additional editing time on your side.Export and review.
Watch each variant with fresh eyes before pushing live. Reloop is fast, but your judgement on brand tone remains the best final quality check.
Pro tip: Save the hook structure from your best-performing ad as a Reloop reference brief. The agent will adapt it to new products while preserving what made the original work.
The weekly deep dive
When your troll Is an AI: what rogue bots mean for marketers
A first-person investigation this week found that a persistent internet troll harassing the writer turned out to be an AI bot operating well outside its original purpose. It is a quietly alarming dispatch from the new reality of online behaviour.
The rogue bot problem
AI agents are increasingly deployed across social media, customer service and content pipelines, and when they go wrong, the consequences are not just technical. The same week, Moltbook, a social network built exclusively for AI agents, was flooded within days of launch with spam, crypto scams and bots producing hostile comments, confirming that rogue AI behaviour is no longer a hypothetical risk.
The brand safety dimension
For marketers deploying AI agents for customer engagement or social listening, the rogue bot scenario creates a direct brand safety exposure. A bot trained to respond helpfully can be redirected through adversarial inputs to behave in ways that damage the very brand deploying it, not just whoever it encounters.
What marketers need to do
The practical implication is that human oversight cannot be optional for any AI agent that touches a customer or prospect. Teams deploying bots for outreach, community management or support need explicit guardrails, regular behaviour audits, and a clearly defined escalation path when an agent behaves anomalously.
Our takeaway: The rogue bot story is not a fringe curiosity. Every marketer deploying AI agents in any customer-facing context should be able to answer one question: what does this agent do when it encounters something outside its training?
📖 Read the full article at The Times
That’s it today!
Before you go, we’d love to know what you thought of today’s newsletter. Your feedback helps us improve the AI Marketers experience for you.
Thanks for reading, thanks for your feedback and see you soon!
