October 23rd-October 29th // Estimated Reading Time: 9 minutes
In This Edition!
Meta controls your targeting
OpenAI controls your creative
You’ve bought AI tools. You’ve built AI products. But your team sounds and acts the same, and the revenue needle isn’t moving. You need to:
a) Build an AI playbook that everyone can use…
b) Enable your entire team on the playbook…
In order to create real, significant business impact.
Leveling up your team on AI is the biggest growth lever you can pull in 2027. Don’t take it from us, take it from BCG. We can help. Let’s chat!
Top Headlines🔥
Meta’s enhanced AI ad ranking system makes targeting less important
Source: AdExchanger
October 28th, 2026
Summary: Meta’s ad system today works as follows: “Andromeda” narrows tens of millions of ads to a shortlist for each user, then a ranking model picks the winner. Meta rolled out three upgrades to this system. Upgrade one is GenRecs, which uses LLMs to describe every ad by traits like brand, color and style, so Meta can match new ads to people based on their preferences and past behaviors. GenRecs also powers a new retrieval model that adds better candidates to Andromeda’s shortlist. Upgrade two is a faster Andromeda, rebuilt in partnership with Nvidia, that now scores more than a billion ads per second. Upgrade three is a bigger version of GEM, Meta’s largest ads model, which now reads twice as much of each user’s history. Meta says its ranking improvements drove 8.3% more clicks and 15.7% more conversions on Facebook in Q2. Advertisers don’t need to change any settings, and Meta says human-set targeting matters less as its AI models take over audience selection.
Opinion: Meta’s AI uses an ad’s brand, color, and style to match it to a user. So … your creative is now your targeting. If every one of your ads uses the same palette and look, Meta will keep matching you to the same people, which makes strict brand guidelines a limit on your reach. Your takeaway: keep the message consistent and vary the look. And create lots of variants.
OpenAI will now rewrite your ChatGPT ads unless you opt out
Source: AI Ready CMO
October 26th, 2026
Summary: OpenAI can now rewrite ChatGPT ads in real-time to match the conversation they appear in. It can change an ad’s headline and description, translate it, resize images, and send clicks to different pages on the advertiser’s site. These real-time variants can run without the advertiser’s approval. OpenAI’s help page says advertisers won’t get a preview of every potential version. The help page also doesn’t indicate that there’s any reporting that shows which version ran. The change applies to campaigns bought through OpenAI’s self-serve Ads Manager and was scheduled to start October 29th or 30th, depending on which OpenAI source you read. Advertisers can opt out of the technology. OpenAI recently explained how ChatGPT chooses ads: Instead of reacting to one prompt, its AI looks at the whole conversation to figure out what to show.
Opinion: Meta looks at your ad to pick the audience. OpenAI reads the audience to rewrite your ad. Same idea: the platform decides what the ad says and who sees it. So what does the advertiser still control? The offer, the product, the landing pages. Put your energy there.
New Products & Features 🚀
What It Does: Google calls Gemini 4 Carbon its most capable AI model, released two weeks after Gemini 4 Argon. Carbon is better than Argon at cybersecurity, makes about half as many unsupported claims, and can work for days on hard coding problems. Only select cybersecurity teams can use it for now. Another new model from Google, Genie 4 creates interactive simulated worlds that people can build using natural language. Google pitches it as a creative tool for filmmaking and gaming, which could help brands build interactive experiences.
Quick Take: Branded game worlds on Roblox or Fortnite typically require a big budget and a specialist studio. Genie 4 significantly lowers the barrier to entry for brands to create interactive experiences and games.
What It Does: Bluefish started as an AI visibility platform that helps brands track and improve how they show up in AI. Its new Paid Media Briefs tool uses that data to find where a brand is missing, misrepresented, or already recommended, but sub optimally. It then suggests targeting, creative, and landing pages to address those gaps. The goal is to manage organic and paid AI marketing in one place, instead of with separate teams and tools. Bluefish plans to add retail media activation, creator partnerships, and sponsored publisher content next.
Quick Take: We said this last week about Evertune’s new Starter plan: As AI visibility data becomes more of a commodity, it makes sense for GEO (generative engine optimization) platforms to move toward paid ads.
Same applies here. Expect every GEO platform to aggressively move toward paid ads.
What It Does: Tatari’s new AI-powered tool scores each ad slot on how well it fits a brand, based on how viewers responded to ads in the past. Then it picks which brand’s ad runs, in real time. Tatari Max works across both CTV and linear for ads bought directly from networks like NBCUniversal, Disney, and Tubi.
Quick Take: It’s like Performance Max, but for TV! Smart product idea. Makes sense. Might be difficult to execute because of the nuances of CTV and linear TV advertising, but the promise is there. That said, please, for the love of god, no more Max and Plus products, please. We’re begging you.
AI Use Case of The Week💡
Clorox trades big marketing campaigns for small, fast AI-powered tests
The Setup: Clorox used to plan one major campaign at a time, months in advance, and run it for a year. Each one needed sign-off from about a dozen decision-makers. CMO Eric Schwartz said that pace can’t keep up with how fast online culture moves now.
The AI Solution: Clorox now runs what it calls “systems of continuous relevance.” AI helps teams detect signals, make bets, and adjust while content is live. A tool that scans product reviews🔒 shows what customers care about most. Hidden Valley uses AI to generate ads, pairing the dressing with different foods. Clorox Pure, a new line of sprays, uses AI and in-market data to help update live ads without a reshoot. Pine-Sol and Brita now use AI to post high volumes of social content and let engagement guide what they’ll post next.
The Results:
• Pine-Sol averages 150+ organic posts per month with a 13% engagement rate
• Four limited-edition Pine-Sol products sold out on TikTok Shop within hours
• Brita’s organic impressions grew 1,742% year over year
• A Brita song and animation hit 10M views without any ad spend
Why This Matters: AI is helping Clorox shrink the time between spotting a signal and acting on it. Most brands treat ad campaigns as finished products, but Clorox is treating them as drafts that will be iterated.
Your Action: Clorox’s old model required a dozen sign-offs. Find out how much of your approval chain AI could handle. List everyone who signs off on your ads and what each person checks: brand, legal, product claims, regional. Paste that list, your brand guidelines, and your approved-claims doc into AI and ask: “Turn each approver’s job into a written checklist. Mark which checks are rules you could enforce and which need human judgment.” Then run your last five rejected ads through the checklist. If the AI catches most of what your reviewers caught, you have a case for letting small tests ship without a human-approval process. Humans can approve the big bets.
Other Notable Headlines📌
Google confirms AI-generated review summaries in Search ads - Google’s AI will pull reviews from across the internet and add short, positive summaries to search ads.
Anthropic’s founders pledge to keep Claude ad-free as the company files for its IPO🔒 - The founders said ad revenue tends to push companies to chase engagement, which corrupts AI’s value prop. They also warned they may slow down to make Claude safer, even if it hurts their competitive position.
Nvidia commits $3B to Meta’s Muse - The investment is part of Meta’s plan to raise $95B for AI, and it will go toward making Muse models faster to train and run.
Amazon hedges against AI shopping agents with 250M outside listings🔒 - Amazon still blocks outside AI agents like Meta’s Muse from shopping on its site. At the same time, its Shop Direct program now shows products sold on other companies’ websites, including brands that don’t sell on Amazon, like Alo Yoga and Warby Parker.
Publicis gives its AI platform Marcel its own CEO - Marcel started in 2017 as an internal tool to connect Publicis employees. It now builds AI agents for clients using Publicis’ data. New CEO Scott Hagedorn will expand it into creator marketing, sports, and AI visibility.
AI is eating into Google's search revenue - Search revenue grew 12% to about $63B in Q3, missing estimates for the second straight quarter, yet search query growth hit a five-year high. Analysts worry AI chatbots and agents are taking share.
Cindy Rose on how AI agents could boost margins🔒 - The WPP CEO is pushing clients to pay for results instead of staff hours. That would let WPP use agents for tasks like versioning, translation, and research, which cost less than people, boosting margins.
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, ad tech, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!
Thanks for reading!
