January 16th-January 22nd // Estimated Reading Time: 8 minutes
In This Edition!
Ads in ChatGPT! They’re here!
ChatGPT wants its cut of commerce
Check out Ep. 8 of the AI Edge Podcast with special guest Mike Ditter, Director of AI Strategy at Diageo. We talk AI innovation at CES, Mike’s pro tips for closing the AI human capability gap within a large marketing organization, and … how much dinosaur pee is in your drinking water.
Listen on
🍎Apple
🌐Spotify
#KnowledgeIsPower #HumansAheadOfAI
Top Headlines🔥
OpenAI may charge marketers $60 CPMs to test ChatGPT’s first ads
Source: Ad Age, The Information
January 20th-21st, 2026
Summary: It’s finally happening! OpenAI is courting advertisers to be the first to test ads in ChatGPT, reportedly at $60 CPMs (for reference, this is what Netflix was asking for when it launched streaming ads in 2022). OpenAI is asking advertisers for “under-$1M spend commitments” over a trial period, with ads launching in early February. ChatGPT will charge based on impressions rather than clicks, and ads will appear only to US users of the free version and new $8/month ChatGPT Go tier. With 900M weekly users, OpenAI is positioning this as "premium" inventory, though only 24% of ChatGPT queries have commercial intent, or users seeking information about products. Insiders say retail and travel brands with strong product data feeds that already optimize for AI visibility could be ideal for early testing. Ads won't influence ChatGPT's organic answers, but the organic conversation will determine which ads get shown.
Here are OpenAI’s “Ad Principles”:
In response to the news, Google said it wouldn't feature ads in Gemini, its ChatGPT competitor. Okay, Google.
Opinion: The $60 CPM price tag is for shits and giggles. OpenAI is positioning its ads product as “premium” to start in order to set the bar high and learn from blue-chip advertisers, but this will become a performance channel fast. It has to. That's the only way to achieve meaningful scale quickly and raise as much money as it wants to in order fund its infrastructure ambitions.
The big question for marketers: how will you buy these ads? Early tests will probably mirror search advertising and thus will favor brands with strong product feeds (think retail, travel, CPG). But unlike search where you just bid on keywords, AI chat will eventually require a fundamentally different approach—you're not going to be buying one-dimensional, in-the-moment, intent signals, you're going to be buying contextual relevance within ongoing conversations over time. Figuring out the right buying mechanisms in order to make sure ads are performant will be critical for OpenAI and its advertisers.
Smart brands should be testing and scenario planning now: How will targeting work? What kind of data will be available? What levers will be available for optimization? What should success look like? How much budget should an advertiser carve out for testing without sacrificing proven channels?
Every LLM will rush to launch ads now that ChatGPT has paved the way. They’ll all be walled gardens. They won’t coordinate or standardize advertising. Measurement will be a mess.
But don't overthink it—just start testing. Even expensive lessons now will be cheaper than playing catch-up later.
ChatGPT Checkouts to take 4% cut of Shopify merchant sales
Source: The Information
January 21st, 2026
Summary: Starting January 26th, Shopify merchants' products will be available for purchase through AI chatbot checkouts including ChatGPT, Google's AI Mode and Gemini, and Microsoft Copilot. OpenAI will charge merchants a 4% fee on sales made through ChatGPT's checkout feature, in addition to Shopify's standard transaction and payment processing fees. Google and Microsoft are not charging additional fees for now. Merchants must opt in to AI platforms that charge fees for checkout, but if they don’t, their products will still appear in AI responses with links back to their websites for checkout.
Opinion: Google and Microsoft offering fee-free checkout creates immediate competitive pressure for OpenAI. But they need to make money anywhere they can. Just goes to show how hard it is to compete against the tech behemoths, even if you have a head start.
Brands should test all three platforms and let the data decide. The fee pays for itself if ChatGPT drives 4%+ more revenue per visitor through better conversion or larger baskets.
New Products & Features 🚀
What It Does: NAVI automatically resolves delivery issues, returns, refunds, and exchanges for retailers, based on the company's policies and real-time context. Powered by Narvar's IRIS intelligence layer, which is informed by 74B consumer touchpoints, NAVI orchestrates post-purchase workflows across systems to handle customer service issues autonomously while protecting margins and brand integrity.
Quick Take: Every dollar spent acquiring customers through marketing is wasted if post-purchase experiences drive them away. NAVI protects that investment by automating resolution of the delivery and returns issues that most commonly erode customer satisfaction and lifetime value.
What It Does: Algolia has integrated its real-time product data, including attributes, inventory, and pricing, directly into Microsoft Copilot, Bing Shopping, and Edge. This lets retailers push their product information into Microsoft's AI shopping tools instead of relying on web crawlers that may surface stale product data. Early pilots with brands like Little Sleepies and Shoe Carnival show how retailer-approved data helps them show up better in AI queries.
Quick Take: AI shopping is now a vehicle for product discovery and conversions, essentially becoming a new performance marketing channel. AI shopping visibility greatly depends on how well your product attributes match natural-language queries, making feed quality a direct revenue driver brands can measure, test, and optimize.
What It Does: Anthropic says Cowork is a simpler way for non-technical people use Claude’s agentic capabilities for everyday tasks. Simply set a task or list of tasks and Claude will plug away. For example, it can turn your scattered notes into the first draft of a report or read, edit, or organize files in folders on your computer if you provide access.
Quick Take: People are going nuts for this tool. We can see how it would make life easier and help marketers create hyper-specialized agents for narrow tasks like "organize campaign assets" or "compile screenshots into a brief." Apparently Claude Code autonomously built Cowork in a week and a half!
AI Use Case of The Week💡
The Setup: Booking.com's TikTok presence generated massive engagement, but as the brand scaled both organic and paid content on the platform, the team faced a growing challenge to accurately analyze thousands of comments to understand sentiment, extract insights, and route responses to the right teams.
The AI Solution: Booking.com implemented Sprinklr’s AI-powered comment moderation to automatically detect intent and sentiment at scale. The system categorizes comments and routes them to marketing or care teams. It also analyzes comment sentiment alongside paid content performance, helping Booking.com identify which boosted posts resonate and adjust media spend accordingly.
The Results:
• 9,500+ TikTok comments analyzed over 60 days
• 2,000 comments identified as requiring engagement and routed to appropriate teams
• 17+ hours of manual work saved
Why This Matters: Comment sections can act as performance data streams. Brands treating comments solely as support tickets miss critical signals about creative effectiveness and audience fit. As brands use AI to scale up their content production, automated sentiment analysis could help them improve quality and performance.
Your Action: Explore how positive comment density (such as positive comments per 1,000 impressions) can act as an early quality indicator for your TikTok content, especially if AI tools are used to create it. Set a “positive comment floor”—for example, content must generate at least 5 positive comments per 1,000 impressions before you boost it with paid spend.
Other Notable Headlines📌
IAB's new framework helps marketers avoid AI disclosure overload - The risk-based guidelines require advertisers to disclose that they’re using AI only when it materially affects authenticity (synthetic voices, digital twins, AI-generated events) but skip disclosure for routine AI uses like editing and optimization.
Amazon may let other companies run their chatbots on its site - In an interview this week, CEO Andy Jassy confirmed that the retail giant is considering how "to partner with other companies to run their chatbots on its site." Yes, AI is going to buy things for us.
Brands are struggling with 'click-less journeys’ - Sixty-two percent of marketers say AI-driven click-less journeys will be standard within a year, but only 27% feel prepared.
OpenAI is reportedly developing earbuds while Apple works on wearable AI pin - OpenAI is expected to announce AI earbuds in the second half of 2026, while Apple's pin was described as a thin, flat circular disc about the size of an AirTag that includes two cameras, three microphones, a speaker, and a Fitbit-like charging strip.
Remember that other AI pin? Humane? That didn’t go so well…
That’s It For This Week 👋
The U of Digital AI Newsletter is intended for subscribers, but occasional forwarding is okay!
To subscribe visit Uof.Digital/Newsletters 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, ad tech, mar tech, privacy / identity, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!
Thanks for reading!
