September 4th-September 10th // Estimated Reading Time: 7 minutes
In This Edition!
You can now buy toilet paper AND ChatGPT ads through Amazon
Can Meta make agentic AI mainstream?
Top Headlines🔥
Amazon's advertisers can now buy ads in ChatGPT
Source: CNBC
September 10, 2026
Summary: Advertisers that buy ads through the Amazon DSP can now extend their campaigns to ChatGPT. Just like they could through Criteo, StackAdapt, and other ad platforms. The campaigns can be set up and managed through Amazon (e.g. budgeting, bidding, creative) but OpenAI's system will control ad decisioning. Delta Vacations is among the first brands testing it. The deal comes after Amazon became the top retail advertiser on ChatGPT between April and August, according to Sensor Tower, even as it restricts ChatGPT and other AI platforms from accessing its own web site and product catalog.
Opinion: Seems like a win-win-win. Advertisers that use the Amazon DSP get access to ChatGPT inventory without having to use a new platform, Amazon expands its ad supply, and OpenAI gets access to Amazon’s ad demand. Amazon's ad demand, like Criteo’s, is heavy on retail and CPG, which should theoretically be a great fit for AI.
OpenAI is desperate to show hockey stick ad revenue growth, given its upcoming IPO. So while the company might aspire to become a walled garden to rival the likes of Amazon and Google, it doesn't have a choice but to tap into the ad demand from these platforms. At least right now. A similar ad partnership between OpenAI and Google would make a lot of business sense for both sides, but Gemini vs. ChatGPT competitive egos likely prevent such a deal from ever happening.
Marketecture Live Chicago Is On September 23rd!
Get Your Ticket Now
New Products & Features 🚀
What It Does: Meta's new AI agent handles real tasks instead of just answering questions, like booking travel or filling out forms. It can shop and complete purchases on a person's behalf, but it will check with the user before pulling the trigger. Muse can also remember details like saved recipes and dietary restrictions to make suggestions unprompted. Muse won't share conversations or data with Meta's ad systems (for now).
Quick Take: While prompt-based AI has become mainstream, agentic AI hasn't. Eventually, it will. If Meta can use its distribution to get traction on Muse with consumers, it has a chance to leapfrog ChatGPT, Claude, and Gemini. And because Meta is already an ad powerhouse, ad monetization won't be hard.
What It Does: The blueprint gives engineering teams the code, guardrails, and other information needed to get a shopping or merchant agent running on Claude. The shopping agent can search a catalog, build a cart, and answer customer service questions in one conversation. The merchant agent can analyze sales data and draft pricing, promotions, and marketing campaigns. Anthropic says retailers running shopping agents on Claude have seen carts up to 35% larger and a 60% higher chance shoppers complete a purchase.
Quick Take: Enabling commerce opens up all sorts of ad monetization opportunities for any AI platform. Hard to believe that Anthropic will always be anti-ads.
What It Does: The Commerce Power Pack bundles four tools. Dynamic Product Ads for Sponsored Snaps brings AI-powered catalog product ads. Value Optimization optimizes campaigns for purchase value. Third-Party Web Optimization combines first- and third-party signals to optimize campaigns toward advertiser KPIs. Omnichannel Optimization helps campaigns capture sales across whichever channel is most likely to convert. Snapchat says Dynamic Product Ads campaigns optimized for purchases saw a 22% improvement in 7-day cost per acquisition.
Quick Take: AI + Commerce + Ads = 🔥
AI Use Case of The Week💡
L'Oréal used AI to brainstorm new ways of reaching male audiences
The Setup: L'Oréal wanted to reach male consumers for its CeraVe skin care line in Australia and New Zealand, a demographic most personal care brands underinvest in. The brand needed to move quickly and develop the campaign in weeks, not months.
The AI Solution: L'Oréal discovered the AI tool Springboards through its own internal innovation competition. It used the tool to brainstorm taglines and creative directions for the campaign. Springboards' model increases variation at key decision points in a sentence to avoid generic outputs. L'Oréal fed the model consumer insights, brand guidelines, and CeraVe product information, then ran collaborative sessions where the marketing team gave live feedback and refined prompts until they landed on a final tagline.
The Results:
• 40 unique campaign ideas generated in two weeks
• Springboards identified sports content as the way to reach men, timed to Australia's football and rugby playoffs
• Final tagline, "CeraVe, Your Skin Barrier's MVP," adapted from AI-generated options
Why This Matters: L'Oréal used AI to quickly widen the pool of ideas, then applied human judgment to pick and refine the winner. That division of labor (AI for volume and variation, humans for final judgment) kept the output on-brand.
Your Action: Treat AI as a brainstorming partner. Generate a batch of options for a campaign idea or tagline, review them as a team, and use that discussion to refine the prompt before generating again. Repeat that loop a few times, and don’t forget to feed the model your brand guidelines, audience data, and other proprietary intel. The human judgment at the end, including picking, combining, and polishing the AI's raw output, can hopefully get you closer to ideas that work well and sound like your brand.
Other Notable Headlines📌
Historic NYT v. OpenAI copyright battle heats up - A win for the Times in a lawsuit over AI training data could upend the business models AI companies depend on. A win for OpenAI and Microsoft would weaken publishers' ability to control and monetize how their content gets used.
Advertisers want more brand safety, prompt-matching, incrementality tracking from ChatGPT ads🔒 - Ad buyers say ChatGPT ads remain a reporting "black box," with no visibility into the prompts triggering ads or negative keyword controls like they have on Google Search.
Magnite runs its first agentic CTV campaign in EMEA - Magnite's AI buyer and seller agents built and activated a premium CTV campaign for an automotive brand using natural language prompts. It cut campaign setup time by about 70% and hit a 95% video view-through rate.
OpenAI restricts Adobe and other competitors from advertising in ChatGPT🔒 - OpenAI updated its ad policies to reject campaigns for standalone image- and audio-generation products.
DeepIntent releases agentic ad platform as pharma marketing grows more complex🔒 - DeepIntent, the pharma-focused DSP, launched ‘Cora’, which runs pharma ad campaigns end to end, from audience strategy to media execution. In beta testing, it cut planning time by 50%.
Actionable raises $10M to build a better understanding of cause and effect in customer service - The French AI-driven platform predicts customer churn before it happens. It will use the funding to expand to the US.
OpenAI CFO Sarah Friar says ChatGPT's ads should be like 'if Google and Meta had a baby' -🔒 Friar described the vision as combining Google's high-intent search behavior with Meta's personal context on social media. We get what she’s saying, but for some reason, we’re not that excited about it…
That’s It For This Week 👋
The U of Digital AI Edge Newsletter is intended for AI Edge Pro subscribers, but occasional forwarding is okay!
To subscribe visit Uof.Digital/Newsletters/ai-edge or AI Edge Pro, or contact us directly for group subscriptions.
And remember, U of Digital helps teams drive better outcomes through structured education on critical topics in marketing / advertising like AI, mar tech, programmatic, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!
Thanks for reading!
