How Coca-Cola’s Internal Thinkers Are Talking About AI

Why Coca-Cola and AI matter

In 2023, at the white heat of AI hype, Coca-Cola was already talking about some of its experiments in the space and where it would go next, but Global CMO Manolo Arroyo was also alive to the changes it would trigger in working culture, and the need for training.

A year later, the conversation was shifting to effectiveness not just efficiency and, with that, responsibility: “I think a company and a brand like us has a moral obligation to ensure that we really [understand] how is the right way to use AI into the future,” he told a Cannes audience. More recently, the company hit the headlines with an AI-generated Christmas ad, which some speculated may have been more about demonstrating capability than driving towards something new. Despite the AI concerns from creatives, it’s still likely that the company’s wide deployment of its potent brand assets will still deliver.

But this is simultaneously about a story bigger and more mundane. Enterprise software, which is what a lot of this boils down to, is rarely a brisk installation. Investments and plans made now will need to be relevant and useful in five years’ time, which requires wide-ranging scenario planning as well as the role of the human worker or consumer in a world potentially mediated by algorithmic intelligence.

AI design for the future

Speaking at Most Contagious London, Dominik Heinrich, The Coca-Cola Company’s global head of AI design – who was clear in saying that the views expressed were only his, not the brand’s – put it simply: “it’s not a hype.”

It’s not a simple plug and play, however, even for a technology that Heinrich believes will one day become as “invisible as electricity” or the internet.

He breaks down the phases of adoption into three steps, with the current state of AI somewhere in the foothills of the first wave:

1. Save time, cost, efficiency – the phase we are in today. Businesses are streamlining operations, cutting costs and increasing efficiency.
2. Quality, better output – using AI to enhance quality. This is where AI isn’t just making processes faster but better.
3. New systems, transformation – the final wave is something completely new. It’s about creating radically new systems.

The changes will be cultural, and even at the level of the words we use we ought to be more careful. For instance, artificial intelligence calls up too much baggage, “we should really call it algorithmic intelligence,” he says, noting the lack of and hinting at the contradiction in terms of an artificial intelligence capable of consciousness.

Our work should be instead to think about inputs, given that is what it works with: “[the machine] takes your input and predicts the most likely output”, he says. They aren’t thinking by themselves – they just wait to see what we will ask of them. Which also comes down to the fact that the most likely output is a pleasing, often uniform output: “if you’re not asking critically, they won’t answer differently”.

Given that we’re in the “age of mediocre”, AI’s real use should be to allow more time for the kind of slow, critical thinking: making connections, going beyond the surface, that creative humans should be doing.

Where this goes

Heinrich envisages a handful of possible futures, among them interfaces that adapt to the user; omni-modality of the technology as it moves out of computers and phones and into other devices capable of interacting not only to text and sound but other senses. Ultimately, it all comes down to a vision of AI as truly invisible, integrated into contact lenses or other appendages that “optimise you”, everything mediated by an AI layer. At best, this means no more reading emails; at worst, who knows?

The vision, then, is that brand choice is likely to be entirely delegated to an AI, as is likely the brand communication, but Heinrich notes that while the choice can be made by a machine, the ultimate experience of a brand’s product or service cannot. The result, he believes, will be more emotional advertising to humans rather than advertising rational benefits – that side will fall to machines.

However, from the what-to-do-now standpoint, there’s a deeper truth: it’s unlikely that adopting AI will lead to competitive advantage. Instead, companies ought to shift from functions to outcomes, deploying critical thinking and creativity throughout in service of making the ultimate human experience of a user or buyer better. Serving the individual, he says, will be the fundamental goal.

“Mass market won’t be a topic anymore, instead it will be niche at scale.”