Welcome back to the part of my week I look forward to most, the one where I read the ad industry's press releases so you don't have to, then say the quiet part into a microphone. Here's the thread stitching every story together this week, whether the people pitching them know it or not: the machines are taking over the buying, the selling, the troubleshooting, and apparently the deciding when you're hungry. WPP wants to write the rulebook for how robot buyers talk to robot sellers, which is a generous way of saying it wants to own the casino and call it a "standards body." Hyundai decided it would rather own its brain than rent Google's, which is the most self-aware thing anyone has said in advertising since "trust us." OpenAI just poached one of the actual humans who invented the technology making all this possible, because nothing says "we've peaked" like billionaires fighting over the guy who wrote the math. And Papa Johns is staring into your refrigerator like a concerned ex who "just wants to make sure you're eating." Different headlines, same movie. We spent a decade screaming that programmatic was too automated and too opaque, and the industry heard us, nodded thoughtfully, and built something even more automated and gave it a chatbot personality.

What ties it all together is the new religion nobody will admit to on the Croisette this year: intelligence itself is the product now. Not the ad. Not the platform. Not the data. The model doing the thinking. Everyone is suddenly deeply, spiritually concerned with who owns the algorithm, who hoards the talent, and who gets to set the rules before anyone notices the rules are the entire game. The buzzword bonfire will be magnificent, and somewhere between the third glass of rosé and a panel titled "The Future of Trust," a handful of companies are quietly building the toll booth every dollar has to drive through. So pour yourself something, set your expectations for the keynotes somewhere around the ocean floor, and let's get into who's actually building the future versus who just expensed a yacht to talk about it.

WPP arrived ahead of Cannes waving the latest industry trophy: a prototype standard for AI-powered media buying agents. The company has assembled an unusually broad coalition of publishers, platforms, and standards bodies to define how automated buyer agents and seller agents should communicate, transact, and make decisions. On the surface, it sounds like plumbing. In reality, it's a power play disguised as collaboration. Whoever establishes the rules for agentic media buying gets a head start on shaping how billions in advertising dollars move through the ecosystem.

The funny part is that the industry spent years complaining that programmatic advertising became too opaque, too automated, and too dependent on black boxes. Now everyone is sprinting toward a future where machines negotiate with machines while humans monitor dashboards explaining what the machines decided. WPP isn't really selling a buyer agent. It's selling the possibility that it can become the Switzerland of AI-driven media transactions before everyone else realizes that's where the money is.

Hyundai's test with Chalice AI and OpenX may sound like another obscure infrastructure announcement, but it's actually one of the more important developments happening beneath programmatic advertising. By moving its bidding intelligence directly into OpenX's environment, Hyundai gains access to significantly more supply, richer data, and faster decision-making than traditional DSP workflows allow. The result is less waste, better targeting, and a direct line to the signals that matter most.

The real headline is Hyundai's insistence on owning its own decision-making intelligence. While Google and other major platforms push proprietary AI bidding systems, Hyundai's leadership openly questioned whether brands gain any true advantage from using the same algorithms as everyone else. Owning the model, not renting the model, is becoming the new strategic battleground. In a world flooded with AI tools, the companies that build unique intelligence may outperform those merely leasing someone else's.

OpenAI scored another major victory in the escalating AI talent war by hiring Noam Shazeer, one of the architects behind the Transformer technology that powers modern generative AI. Shazeer helped create foundational breakthroughs that enabled today's large language models and recently returned to Google after the company's expensive reacquisition of Character.AI. Now he's leaving again, this time for OpenAI.

The move underscores a growing reality in artificial intelligence: the biggest competitive advantage isn't infrastructure, capital, or products—it's people. The industry's most influential researchers have become the equivalent of franchise quarterbacks with PhDs, commanding enormous compensation packages and reshaping entire company trajectories with a single career move. OpenAI keeps proving that while everyone talks about compute, talent remains the ultimate scarce resource.

Yahoo Finance launched a dedicated media and advertising hub featuring content from several major trade publications, hoping to give investors a clearer view into the increasingly complicated advertising marketplace. The initiative arrives as media, AI, streaming, retail media, and ad-tech become deeply intertwined investment stories rather than niche industry topics.

This is a sensible move disguised as a content partnership. Investors are constantly trying to understand how advertising trends affect platform valuations, media companies, AI investments, and consumer behavior. Meanwhile, advertising executives are eager for Wall Street to believe that their latest buzzword-heavy presentations actually translate into growth. Bringing those audiences together creates value for everyone involved, especially if it helps translate ad-tech jargon into something resembling English.

Google Ad Manager unveiled Ask Ad Manager, an AI-powered assistant designed to help publishers troubleshoot campaigns, diagnose delivery problems, analyze bidder performance, and navigate the platform more efficiently. The chatbot aims to replace hours of manual investigation, endless support tickets, and spreadsheet archaeology with conversational prompts and automated recommendations.

The product exists because ad operations has quietly become one of the least glamorous yet most frustrating jobs in digital advertising. Every day involves tracking down mysterious delivery issues, decoding reporting anomalies, and navigating interfaces that often feel frozen in time. If Ask Ad Manager can eliminate even a fraction of that pain, publishers will embrace it quickly. The only unanswered question is how often the chatbot will confidently invent solutions that don't actually exist, a feature the AI industry continues to market as a bug.

Papa Johns partnered with NBCUniversal, Instacart, and Carat to create a campaign that predicts when consumers are likely running low on groceries and serves streaming ads at exactly that moment. By analyzing grocery purchasing behavior, the company hopes to identify the sweet spot where convenience beats meal planning and pizza becomes the obvious answer.

The campaign highlights both the promise and the creepiness of modern advertising. Marketers are no longer targeting demographics or interests. They're targeting moments of vulnerability, convenience, and impulse. Your empty refrigerator has become a media signal. The industry's challenge is no longer gathering data; it's figuring out how to use increasingly intimate behavioral insights without triggering consumers' internal alarm bells. Papa Johns may have found that line. Or it may have stepped directly on it. Either way, every marketer at Cannes will be pretending they thought of it first.

There are two things Papa John's reliably gets wrong. One is the apostrophe in its own name. The other is the part where it markets to you without thinking through how that looks in a courtroom.

This month the chain announced a campaign it is calling, with the subtlety of a pizza box left in the sun, "Empty Fridge." The setup: Papa John's hired NBCUniversal, Instacart, and the dentsu agency Carat to figure out the precise moment you are out of groceries, tired, and weak. Then it serves you a streaming ad on Peacock or NBC Sports with a QR code, so you can summon a large pepperoni without ever rising from the couch you have melted into. Instacart builds the audience by watching who keeps buying eggs, milk, meat, and produce, then models the day of the week each household is likely to hit empty. The ad arrives at the bottom of the cycle.

That is the whole idea. That is the pitch they brought to clients with pride.

Carat's chief investment officer, Carrie Drinkwater, described the goal as knowing what is in your fridge "without being too creepy." I want to sit with that phrase, because it is doing an enormous amount of load-bearing work.

"Without being too creepy" is not a strategy. It is a confession with a smile on it. Nobody says "without being too creepy" about a thing that is not creepy. You do not describe a salad as "barely poisonous."

Let me be fair to the engineers, because they deserve better than the copywriters. The underlying technique is boring and legitimate. Retailers have modeled replenishment cycles for years. Your diapers, your dog food, your coffee pods: there is a known interval, and a propensity model predicts when you run low. Pipe that into a connected TV deal and schedule the ad around it, and you have what the industry calls "moment targeting" and what your mother would call "good timing."

What you do not have is a camera in the Frigidaire. The honest version of the claim is "this household is probably toward the end of its grocery week sometime around Thursday." The marketed version is "we know your fridge is empty at 7:32 p.m." One of those is a model. The other is a magic trick performed for an audience of brand managers in Cannes who applaud anything with a confidence interval. And it only works at all for people who buy groceries on Instacart with enough history to model. For everyone else watching Peacock, there is no fridge science. There is just an ad for pizza, same as it ever was, wearing a lab coat it bought online.

Papa John's CMO Jenna Bromberg was at least refreshingly blunt about the point of the exercise. "Order growth is our No. 1 KPI," she said. Not customer trust. Not "responsible data use." Order growth. I appreciate honesty in a press cycle, even when the honesty is the problem.

The campaign's actual innovation is not the fridge. It is the targeting philosophy. Marketers used to chase demographics and interests. This chases a state of mind. The whole premise is to reach you when you are "running out of steam at the end of the week," which is corporate for tired, depleted, and disinclined to make a decision you would defend while sober. That is not an interest segment. That is a vulnerability segment. The dark-patterns crowd has spent a decade arguing that designing for the moment of least resistance is manipulation. Papa John's just built a media plan around it and put it on a sizzle reel.

And it confirmed something else worth flagging: Instacart says this is the first time it has let a brand outside the packaged-goods world use its first-party shopping data on NBCU inventory. So your grocery history, the data you handed over so your order would arrive correctly, is now behavioral telemetry powering a pizza company's hunt for the moment you are too tired to cook. Nobody clicked "yes, use my egg purchases to ambush me on Sundays." That consent does not exist. It was assumed, the way it always is.

Which brings us to the word the user buried in the brief, and the word I would print in 48-point type: again.

Because here is what happens when regulators decide to take another look at Papa John's, and they will, because they have before. This is the company that paid $16.5 million in 2013 to settle a class action over roughly 500,000 unsolicited marketing texts, a case that at one point carried exposure north of $250 million under the Telephone Consumer Protection Act. The company's defense back then was that you consented to the marketing when you handed over your phone number to order a pizza. The courts did not love that theory. And it kept happening: fresh TCPA suits in 2020, more in 2023, customers swearing they hit "stop" and got texted anyway from a new number weeks later. The throughline across fifteen years is not a glitch. It is a corporate reflex. Papa John's keeps treating "you gave us your data for one reason" as "you gave us your data for every reason."

Now overlay that track record onto a campaign whose entire selling point is inferring a private fact about your home and acting on it at your weakest moment. California's privacy enforcers are already circling automated decision-making and the secondary use of consumer data. The DELETE Act exists precisely because regulators decided shopping behavior should not become a perpetual leash. A company with Papa John's history does not want to be the case study. It wants to be the asterisk in someone else's case study.

So enjoy the QR code. Scan it while you are tired. The pizza is real and probably fine. Just understand that the genuinely innovative thing here is not that Papa John's learned to read your fridge. It is that, after fifteen years and eight figures in settlements, it learned absolutely nothing about the difference between data you offered and data you would have guarded if anyone had bothered to ask.

The fridge is empty. The legal exposure is not.

I keep imagining the pitch meeting, because somebody stood up in front of a client and said the words "we will know when their fridge is empty" out loud, on purpose, and got a budget for it. There was a deck. There were slides. Somewhere there is a person whose LinkedIn now says "Spearheaded cross-platform vulnerability-moment activation" and who genuinely believes this is the good kind of resume line.

And nobody in that room raised a hand to ask the only question that matters, which is: what does this look like to a juror? Because juries do not speak fluent ad-tech. You cannot stand in a Kentucky courtroom and explain "recency-frequency replenishment-cycle propensity modeling" to twelve people who order Domino's. You can only say what you did, which is: we built a machine to figure out when families were too tired to feed themselves, and then we sold them pizza. That sentence does not improve under cross-examination. That sentence gets worse every single time you read it.

Here is the detail that brings me genuine joy. The entire campaign hinges on a QR code, the technology we all collectively agreed was dead in 2014 and then reanimated during a pandemic so we could read menus without touching them. Carat's own investment chief admitted QR codes have been "a little wonky" in advertising, which is agency-speak for they do not work.

So the plan, start to finish, is this. Identify the precise moment a human being is at their most depleted. Deploy fifteen years of grocery surveillance to pinpoint it. Time the creative with surgical precision. And then ask that exhausted, hungry, end-of-week person to pick up their phone, open the camera, hold it steady at the television, wait for the little box to focus, and scan a code before the will to live the convenient life passes. If they are too tired to make pasta, they are too tired for augmented-reality calisthenics with the remote in one hand. The funnel collapses on its own premise. You cannot target laziness and then assign it homework.

If a brand came to me with this, and brands have come to me with worse, I would say: the model is fine, the framing is a subpoena waiting for a stamp. Target heavy pizza buyers on a Thursday. Call it good timing. Win an award nobody investigates. The exact same media buy, minus the part where you go on the record bragging that you can see inside someone's refrigerator.

But that is the boring version, and boring does not get you to Cannes. Cannes wants the camera-in-the-fridge story. Cannes has never once been deposed.

So here we are. A pizza company with a documented, fifteen-year, eight-figure habit of using customer data in ways courts described as "not consent," now publicly proud of a system that infers when you are weak and acts on it. They named it "Empty Fridge." They could have named it almost anything. They chose the one that doubles as the plaintiff's exhibit label.

I genuinely wish them luck. The pizza, again, is fine. It is the discovery process I'd be losing sleep over.

The full ADOTAT Scorecards. Vendor comparisons that name winners and losers, with the analysis your planning team needs before Q3 budget meetings.

The takes nobody else will publish. Independent means we don't answer to vendors, agencies, or the trade press establishment.

The archive. Every Scorecard, every analysis, searchable. The reference layer for how this industry actually works.

Read by senior leaders at every major holding company agency, GroupM, Publicis, Omnicom, IPG, Havas, Dentsu, and inside the buying and strategy teams at Netflix, Roku, Pinterest, Disney, NBCUniversal, Spotify, Amazon, Meta, and most of the platforms shaping where ad dollars actually go.

If your competitors are reading ADOTAT+ and you aren't, you're planning with less information than they are.

$99/month. $999/year (save $189). Expense it.

Founding member rate. Locks in for life.

Also, no ads. Zero ads. No Ads.