Coca-Cola’s Mourinho AI Clone and the Advertiser Blind Spot

Coca-Cola’s FIFA World Cup 2026 campaign pairs José Mourinho, the incoming Real Madrid manager, with an AI-generated digital twin for a daily debate series. More than 200 pieces of content are targeted across five platforms during the tournament, published in English, Spanish, and Portuguese with subtitles in 10 additional languages. Behind each video sits a human production chain: representatives from Mourinho’s own camp watch matches alongside Coca-Cola and Footballco’s editorial team, script post-match reactions, route them through brand and talent approvals, then push cleared material through Google Cloud before the AI-generated output goes live.

The campaign produces daily reactive content timed to match results, a volume traditional celebrity contracts have never delivered. It launches as research from the Interactive Advertising Bureau (IAB), the US digital advertising trade association, shows the gap between what advertisers believe younger consumers feel about AI-generated ads and what those consumers actually report has grown to 37 points and kept widening for two straight years.

A Human War Room Behind the Clone

The series, announced June 4 by Coca-Cola and Footballco, is called “José vs. Mourinho” and will run across GOAL, Footballco’s global football media brand, for the duration of the tournament. Two virtual versions of Mourinho debate opposite sides of the same football topics in real time, combining what Coca-Cola describes as his competitive spirit and sense of humor in a social-first format. The official press release from Coca-Cola’s launch of the José vs. Mourinho series is explicit on this point: entertainment, not serious professional commentary.

The project was built by Footballco’s editorial team alongside GRAiL, a next-generation entertainment studio and management company, using Google Cloud technologies. Mourinho participated directly in its construction: a Coca-Cola spokesperson confirmed to Digiday the team worked with Mourinho to capture new audio and video footage, providing the raw material for a replica that mirrors his mannerisms and communication style. That consent-based capture process is the structural feature that separates this model from unauthorized AI-generated likenesses of talent.

The ongoing production workflow is what the press release leaves out. Per Digiday’s reporting, a human war room runs continuously throughout the tournament: scripting teams watch each match, draft post-match reactions, send scripts to Mourinho’s representatives alongside Coca-Cola brand teams for approval, and only then push cleared material through the AI synthesis step. Nothing publishes autonomously. Stefano D’Anna, president of Footballco, said the campaign’s setup phase ran longer than traditional creative execution to build out those safeguards and approval chains, months of work before a single piece went out.

The Scale Math Coca-Cola Is Buying

The economic case for the model is direct. D’Anna told Digiday that traditionally talent can give a brand one or two hours to shoot material that lives statically across media channels. A handful of spots, some cutdowns, a photo library, and the contract ends when the cameras stop. For a 48-team tournament running 104 matches across the United States, Canada, and Mexico, that production window is the wrong unit of measure.

Sarah May Bates, SVP and group creative director and director of AI and innovation at RPA, an advertising and marketing agency, frames the tradeoff in concrete terms: AI tools can reduce hard production costs, including travel and physical shoot logistics, by 40 to 50%. What those savings enable is not fewer deliverables but more of them, and more differentiated ones. “We now have, let’s say, six spots, and every single one can have its own set of variations that are bespoke to whatever audience,” Bates said.

• 200+ pieces of content targeted for the FIFA World Cup 2026 tournament window, per Coca-Cola’s announcement
• 5 platforms (TikTok, Instagram, YouTube, X, Facebook) receiving real-time daily publishing throughout the tournament
• 40-50% estimated reduction in hard production costs from AI-assisted workflows, per Bates at RPA

The catch is front-loaded. The months of capture sessions, pipeline architecture, and sign-off frameworks required before a single clip publishes means this model does not eliminate the production investment. It relocates it to setup, and the total resource outlay before launch runs longer than a traditional campaign.

The Pattern That Started With Snickers

In the summer of 2024, Mars ran a campaign for Snickers built around a different AI version of the same person. Working with agencies T&Pm and Helo, the brand built an authorized AI clone of Mourinho powered by ElevenLabs, Synclabs, and OpenAI’s GPT-4o, designed to generate personalized video responses to fans who submitted their season’s worst football “Own Goals.” The tool launched June 25, 2024. Per Mars’s press release for the Snickers Own Goals campaign, a team of over 70 people built an eight-stage AI pipeline with integrated safeguarding and a “chain of thought” architecture to keep outputs brand-safe and non-templated.

Mourinho’s Coca-Cola series is the second authorized AI campaign by a different global brand within 18 months. A third campaign, from Avocados From Mexico at the Super Bowl this February, ran actor Rob Riggle as a realistic AI avatar for live game predictions and guacamole recipe interactions. Across the three, the shared logic is consistent: a recognizable public figure, explicit participation consent, and AI handling volume and personalization that a human schedule cannot deliver.

Financial terms for none of the three campaigns were disclosed. Mourinho described the Coca-Cola project to Digiday as “the funniest one yet: arguing with myself every day,” framing his own participation as a feature of the entertainment premise rather than a contractual obligation.

What Advertisers Think Consumers Feel

The production efficiencies are real, though the consumer reception is harder to project. Research from the IAB and Sonata Insights, published in January in the IAB’s AI Ad Gap report alongside the industry’s first AI Transparency and Disclosure Framework, found 82% of advertising executives believed Gen Z and Millennial consumers felt positively about AI-generated ads. Only 45% of those consumers actually felt that way, a gap of 37 points that had expanded from 32 points in 2024. The survey covered 505 US Gen Z and Millennial consumers and 104 advertising industry executives, conducted between October 2025 and January 2026.

Among consumers aged 16 to 27, 39% reported negative feelings toward AI advertising, nearly double the 20% among Millennials. Twenty percent of consumers described AI-using brands as “manipulative,” a descriptor that only 10% of industry executives chose. The IAB’s findings also noted consumers were more likely than advertisers to call AI-using brands “unethical” (16% vs 7%).

The value of working with a celebrity would be that people align with those actual humans and what they stand for. A virtual character virtually liking a brand is a ruse and is essentially disposable.

Steve Slivka, chief experience officer and partner at Manifest, a brand experience company, made that argument in an emailed statement to Digiday. Paul Caiozzo, chief creative officer at Tombras, an advertising agency, said his agency’s experience tracks the same concern: younger consumers, Gen Z especially, actively resist noticing AI in creative work. Tombras’s clients had not yet asked for their own digital twins, though the agency was already using AI tools to scale content production in other ways.

The Mourinho campaign carries a structural counterargument to that concern. The transparency is built into the title itself: both “José” and “Mourinho” name the same real person, and the AI element is the explicit premise of the series, not a fine-print disclaimer. The IAB survey, conducted broadly across AI advertising sentiment, cannot distinguish between campaigns where the artificiality is concealed and those where it is the entire concept.

New York’s Disclosure Clock

As that debate runs through the tournament, a statutory layer arrives in parallel. New York’s Synthetic Performer Disclosure Law, signed by Governor Kathy Hochul on December 11, 2025, takes effect on June 9, 2026, two days before the World Cup opens. The law requires advertisers to conspicuously disclose when ads include a “synthetic performer,” defined as a digitally created asset intended to give the impression of a human performance by someone who cannot be identified as a specific natural person. Penalties start at $1,000 for a first violation and reach $5,000 for subsequent ones.

The definition creates a meaningful limit for the Mourinho campaign specifically. Cooley’s analysis of New York’s synthetic performer disclosure requirement notes the statute targets ads presenting what appears to be a real human performance by a fully synthetic, non-identifiable figure, not AI replicas of named celebrities whose identity is the campaign’s premise. Whether the Mourinho series technically triggers the disclosure requirement is an open legal interpretation, but the statute’s framing suggests it was aimed at a different category of content entirely.

The IAB’s voluntary framework, released the same week as its consumer research in January, draws a wider boundary. It explicitly names “digital twins” among content categories requiring consumer-facing labels, with no exception carved out for identifiable celebrities. The practical result for advertisers building AI celebrity campaigns is two simultaneous standards to satisfy: a narrow statutory one and a broader voluntary one whose industry adoption will shape how the practice develops.

The Longer Bet Beyond the World Cup

Coca-Cola’s history with AI creative is consistent. Its AI-recreated “Holidays are Coming” commercials drew public criticism in both 2024 and 2025, and neither round stopped the next campaign. Mickael Vinet, Coca-Cola’s global vice president of sports, music and entertainment marketing and partnerships, told Digiday the company approaches AI “with humility, in a sense of it doesn’t mean it will be right in the first time all the way.” His longer statement was less cautious: digital capabilities including digital twinning will become “central to Coke’s dialogue with talent and rights holders” as the technology matures, and the company is digitalizing its operations across all fronts, from internal workflows to consumer engagement.

The company has structural room for that kind of experiment. Coca-Cola’s partnership with FIFA spans nearly five decades. Its net revenue reached $12.5 billion in Q1 2026, up 12%, per its quarterly financial results. Consumer backlash to AI creative has not measurably affected those figures.

The bet the campaign makes goes beyond the tournament schedule. Vinet’s statement positions AI digital twinning as a permanent addition to talent negotiation, not a one-time promotional tool for a specific event. The World Cup opens June 11 and the 200-plus pieces that follow are Coca-Cola’s working argument against that 37-point gap.