Chili's CMO on the restaurant's cultural comeback

Full analysis

Your draft

Here's what actually happened: a Fortune 500 restaurant CMO went on a podcast and told AI vendors, politely, to come back later. George Felix, CMO of Brinker International, walked through the Chili's turnaround, and buried in the brand-marketing case study is a real signal for anyone selling AI creative or measurement into the QSR and casual-dining world.

The decision this frames for an ad-tech operator: how much of your 2026-2027 pipeline is built on the assumption that big advertisers are ready to buy AI-powered creative and campaign tools right now? This is easy to undo. Nobody has to commit anything today. But it sets a data point on how fast the enterprise buyer actually moves, versus how fast the vendor pitch assumes they will.

The Market Analyst. Felix is a useful proxy, even though one CMO doesn't make a trend. Brinker is public (ticker EAT), reports same-store sales, and Felix is claiming 20 straight quarters of growth. That's a CMO with the credibility to say no and not get second-guessed by his board. When a winning marketer says AI creative "hype outweighs the impact," that lands harder than a struggling one saying it. For ad-tech vendors pitching generative creative into casual dining, this is a longer sales cycle and a higher bar to clear. In plain terms: the customer with money and momentum is the one telling you your product isn't ready.

The Skeptic. Steelman Felix's position, because it's stronger than it looks. He's not anti-AI. He's running it on operations, cutting admin load on restaurant managers, and keeping it out of creative. That's a rational split. The back office has clear before-and-after numbers. Creative doesn't, and the downside of a flat AI ad in a "human business" is real brand damage for uncertain upside. What has to be true for the AI-creative pitch to work here? That the output is measurably better or cheaper at equal quality. Felix looked and couldn't find the proof. Until a vendor shows him the needle moving, "not in a hurry" is the correct answer.

The Operator. Watch what Felix built, not just what he said. He carved out a distinct "Culture Pops" budget line, separate from performance spend, for high-variance cultural swings like the Scranton Office restaurant. That bifurcation matters more than the AI quote. Restaurant advertisers are splitting money into two buckets: measurable direct-response, and brand bets they judge on volume of swings, not per-campaign ROI. If you're an SSP or DSP pitching reach with no cultural or engagement signal, you're fighting for the shrinking measurable bucket while the interesting money moves to activations you can't easily instrument. The Scranton opening ran on Reddit engagement, not a bid graph.

The Customer / End User. Felix reframed loyalty in a way that should reshape how the restaurant category measures performance marketing. "We just want to be in your consideration set. We're going to be a small percentage of that share of stomach." Loyalty as rotation, not exclusivity. That breaks the clean attribution story a lot of measurement vendors sell. If Chili's only ever wins a slice of a rotating consideration set, last-touch conversion credit overstates what any single campaign did. The advertiser knows this. The measurement pitch that assumes it can isolate incremental restaurant visits to one ad is selling a precision the buyer has already decided doesn't exist.

The CFO. Follow the payback logic. Felix applies one test before adopting anything: does it help guests or restaurant employees? Operational AI passes because the savings are countable. Creative AI fails because the return is a guess and the risk is brand. That's disciplined capital allocation, and it's the same logic every enterprise buyer in this category will apply. For vendors, it means the operations use case sells itself and the marketing use case needs a proof-of-ROI case study you probably don't have yet. Sell where the math already closes.

The tensions

Two disagreements worth naming.

First, the Market Analyst treats Felix as a leading signal for the category; the Skeptic treats him as one disciplined buyer with a defensible split that says nothing about the ten CMOs who'll buy AI creative anyway. Is this a category slowing down, or one confident marketer moving at his own pace?

Second, the Operator and the Customer are pointing at the same crack from different sides. The budget is bifurcating into measurable and unmeasurable, and loyalty is rotation not exclusivity. Both erode the value of tightly-attributed performance measurement in this category. If they're right, the measurement pitch that wins here is the one that stops promising clean incrementality and starts pricing in share-of-consideration.

What it hinges on

The whole thing turns on one belief: can an AI-creative vendor show a marketer like Felix a real, measured lift, at equal or better brand quality, versus his human agencies? Right now he says no such proof exists. That's checkable. If a case study lands where a QSR chain publicly credits AI creative with measurable same-store lift, Felix's stance dates fast. If eighteen months pass and the best evidence is still "faster and cheaper" rather than "better and measurable," he was right and the category buys operational AI while creative stays human.

The council leans with Felix on creative, and with the Operator on budgets. The verifiable move for a vendor: stop pitching AI creative on speed and cost into casual dining. Pitch operational AI where the savings are countable, and build the one measurement product that treats restaurant loyalty as rotation share.

Direct impact of this episode on ad-tech plumbing is low. No programmatic, no identity, no CTV infrastructure. The value is the buyer-sentiment read.

Prediction: No top-10 US restaurant chain will publicly credit AI-generated creative with a measurable same-store sales lift in an earnings call or major trade interview before Brinker International's fiscal Q4 2027 earnings (reported August 2027).

Confidence: Medium. The proof bar Felix names is real and nobody has cleared it publicly yet.

Why: Felix, running a chain with 20 straight quarters of growth, went looking for AI implementations "truly moving the needle in a big way" and couldn't find them, so he keeps AI in operations and human hands on creative. That's the whole category's buying logic in miniature: operational AI shows countable savings, creative AI shows speed and cost but not brand-safe measured lift. For a chain to publicly credit AI creative with same-store growth, someone has to solve attribution in a category where the CMO himself says loyalty is rotation, not exclusivity, which makes clean single-campaign lift hard to isolate. The opposite outcome, a big chain putting its name on an AI-creative sales-lift claim, requires both the measurement problem solved and the brand willing to stake its numbers on it, and no vendor has shown that case study yet.

Revisit by 2027-08-31: We're right if no top-10 US restaurant chain (by revenue) publicly attributes measurable same-store sales lift to AI-generated marketing creative in an earnings call or named trade interview by then. We're wrong if any such chain does, naming AI creative as a driver of same-store growth with a figure attached.