
May 22nd-May 28th // Estimated Reading Time: 9 minutes
In This Edition!
OpenAI copies Meta's ad playbook
Meta copies OpenAI’s subscription playbook

U of Digital Was Live At TWO Conferences This Week!

Myles Younger ran an “Agent Of Change | Building Your AI Agent Dream Team” session at the PTTOW! 2026 Summit in Orlando, FL.

Shiv Gupta ran a “How AI Can Be Used Across The Campaign Lifecycle” session at the Humanic <ai/> x Marketing Summit in San Francisco, CA.

Top Headlines🔥
OpenAI turns on cost-per-action ads inside ChatGPT as it races toward IPO🔒
Source: Digiday
May 28th, 2026
Summary: OpenAI now lets advertisers pay based on when a user clicks on their ChatGPT ads, signs up for something, or makes a purchase. The feature was made possible by OpenAI's recently launched pixel, which connects ad exposure to downstream actions. Select advertisers have already received early access invitations, with a June 1st deadline to set up conversion tracking.
OpenAI split its CMO role🔒 in two: Colin Fleming, former CMO of ServiceNow, takes the B2B side, while a consumer-facing CMO is still to be named. Two recent hires round out the commercial buildout: Sam Mulinder, Snap's former head of marketing science for North America, has been brought on to build out measurement, and former Meta exec Archana Joshi joins as the founding member of its Ads Go-To-Market team. All of this comes as OpenAI races toward an IPO expected to value the company at up to $1T.
Opinion: Not only is OpenAI enabling CPC pricing, they’re going one step further and enabling CPA pricing. Of course, they will have to figure out how to make ChatGPT ads actually convert if they want to scale a performance advertising business.

Initially, everyone was worried that ChatGPT was fashioning itself too much of a brand channel; we now worry that it’s pushing too hard into performance. AI collapses the funnel: product awareness, consideration, comparison, and conversion can all happen here. But if CPA pricing is available, every marketer will view ChatGPT ads as a pure-play performance channel and run their campaigns accordingly. Why box yourself in?

Meta to start testing AI subscription services, with cheapest plan at $7.99 a month
Source: CNBC
May 27th, 2026
Summary: Meta is testing paid subscription tiers for its Meta AI app and website, putting it in direct competition with OpenAI, Anthropic, and Google. Meta One Plus ($7.99/month) gives users more capacity for complex requests and additional image and video generation. Meta One Premium ($19.99/month) unlocks even more compute for deeper reasoning on complex tasks, essentially more access to the app's "thinking mode." Testing starts next month in Singapore, Guatemala, and Bolivia. A free tier remains for casual users.

Opinion: While ChatGPT tries to replicate Meta’s performance ad business, Meta is trying to replicate ChatGPT’s subscription AI business. Meta could spend as much as $145B on AI infrastructure this year, and investors are asking hard questions about ROI. Subscriptions are one way to show them that Meta can build entirely new revenue streams from AI. That said, they have a lot of ground to make up against OpenAI, Google, and Anthropic. Has anyone even used the Meta AI app? We haven’t.

New Products & Features 🚀
What It Does: Cadent's new self-service tool lets media buyers construct audiences from individual data signals rather than buying pre-packaged segments. Instead of picking a black-box bundled segment and hoping the underlying data is relevant, buyers choose and weight their own signals, such as TV viewership, browsing behavior, and foot traffic, to build a custom predictive model. Early results include a 5x CPA improvement for a QSR advertiser and 4.6x behavioral lift for a beverage brand.
Quick Take: The ANA found more than half of programmatic spend never reaches its intended audience, and much of that starts with how audiences are built. AI is a great technology to disrupt audience creation by bringing more nuance and granularity to the process. Could be a boon for audience-targeted open web ad campaigns.
What It Does: Opus 4.8 is an incremental upgrade to Anthropic's most capable model, with improvements across coding, reasoning, and agentic tasks. A new "dynamic workflows" feature in Claude Code lets the model spin up hundreds of parallel subagents to tackle large-scale tasks like codebase migrations. Fast mode is also three times cheaper than before. A notable change is honesty: The model is four times less likely than its predecessor to let flaws in its own code pass without flagging them.

Quick Take: Another day, another model!
What It Does: The partnership combines Optimizely's experimentation and personalization platform with Deloitte Digital's implementation and change management expertise. The pitch is that most companies have invested in AI but can't translate it into measurable results because of the organizational change required to make it stick. The two companies have developed an "AI Blueprint for Marketing Leaders" to help brands redesign their content, personalization, and campaign workflows end to end.
Quick Take: Buying a new AI tool or platform is the easy part, but getting an organization to actually change how it works is where transformation stalls. Enter consultancies. Turns out you do need humans after all (at least for now).
What It Does: Gemini Omni can take any combination of text, images, audio, and video as input and generate video as output. The key differentiator is conversational editing: Users can refine a video across multiple turns, changing the environment, style, or specific details without losing continuity from the original scene. Gemini Omni Flash is available now to Google AI Plus, Pro, and Ultra subscribers, and is rolling out for free to YouTube Shorts creators.
Quick Take: The YouTube Shorts integration is smart. Free access to an AI video generation and editing tool, built directly into the platform where short-form video ads run, is a low-friction way to get creators and brands experimenting.
What It Does: Incrementality measurement platform Measured launched an MCP server that connects its incrementality database directly to AI tools like ChatGPT and Claude. Marketers can ask questions about their media performance and get answers drawn from more than 30K incrementality tests across 200+ brands. Example questions: "Did my Meta prospecting campaign affect Amazon sales?" or "Where should I spend my next dollar?"
Quick Take: Companies that have a true data moat, like Measured, have value in the AI era. And they unlock that value even more when they integrate with AI tools. Companies that have a software moat, well…

AI Use Case of The Week💡
Kellanova uses AI to predict which creative will perform before it runs

The Setup: Kellanova, the snack company behind Pringles, Cheez-It, and Pop-Tarts, now part of Mars, had been manually tracking which creative decisions drove growth. Working with data firm Vidmob and MMA Global, the company set out to automate and scale that process using AI.
The AI Solution: Vidmob's AI broke down 443 ads into more than 20K individual creative attributes, such as logo timing, audio overlays, and whether talent appeared in the opening seconds. It identified 19 attributes that most reliably predicted a 3-second view-through rate, which Kellanova uses as a leading indicator of sales. Kellanova then built separate models by product category, since the signals that work for salty snacks differ from those that work for sweet ones. Agencies were briefed on the criteria and held accountable to hitting a minimum score threshold before campaigns went live.
The Results:
• 83% accuracy in forecasting three-second view-through rates
• 2x improvement in creative performance
• 11% increase in ROI
Why This Matters: Creative has always been part art, part science. This approach tips the balance toward science, giving teams a data-backed brief before production starts rather than lessons learned after a campaign runs.
Your Action: Pull your last 12 months of paid social creative performance data and drop it into an AI tool like VidMob's. Ask it to identify which creative elements correlate most strongly with your top KPI, whether that's VTR, CTR, or ROAS. Examples to look for: product visibility, text overlays, opening scene type, talent vs. no talent. You won't have Kellanova's scale, but even a rough pattern analysis can start to replace gut instinct with something more defensible.

Other Notable Headlines📌
Anthropic raises $65B, nears $1T valuation ahead of IPO - The funding round, which may be Anthropic's last before going public, puts its valuation at $965B. The company's annualized revenue topped $47B this month and expects to hit its first operating profit this year.
X is using AI targeting to win back advertisers after a $2B revenue drop - Per SpaceX's IPO filing, ad revenue on X is starting to come back. SpaceX says ads are a core monetization channel for its AI segment, with a stated goal of making ads "feel like content." The filing also revealed that Grok's image and video generation tool produced 2B videos and 10B images per month in Q1. Many expect a SpaceX-Tesla merger and a SpaceX IPO valuation of $1.75T-$2T.
Manus weighs raising $1B to unwind Meta takeover - After Beijing demanded Meta cancel the $2B acquisition, Manus's founders are exploring a buyback funded by outside investors. Much of Manus's tech has already been integrated into Meta's systems, making the unwind extremely complicated.
Anthropic's Project Glasswing has found 10K+ critical software vulnerabilities in one month - Anthropic and 50 partners have scanned critical open-source software and found thousands of high-severity bugs. The bottleneck has now moved from finding vulnerabilities to patching them.
Meet the Meta exec leading its AI workforce transformation🔒 - CTO Andrew Bosworth is overseeing Meta's shift to an AI-first workforce, which included laying off 8K people and reassigning 7K more to AI roles last week. He defended the controversial policy of recording employee keystrokes and mouse clicks to train AI agents, and refusing employees who wanted to opt out.


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 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!
