Contents
Coca-Cola’s Strategic Approach to AI: Balancing Innovation and Practicality
Coca-Cola’s approach to artificial intelligence (AI) is driven by a clear focus on practical outcomes. Neeraj Tolmare, the company’s chief information officer, emphasizes that any AI initiative must demonstrate potential for revenue growth or significant efficiency improvements before it can be fully implemented. This strategy aligns with the broader vision of CEO James Quincey and the leadership team, who prioritize AI projects that deliver tangible results.
For Coca-Cola, scale is essential. As one of the world’s largest beverage companies, the brand operates through an extensive network of 200 bottlers and 950 production facilities, serving 2.2 billion beverages daily. Every AI investment must account for implementation costs and how the technology will integrate across different departments, from software development to sales. Additionally, Tolmare considers how emerging technologies like agentic AI might reshape workflows and data sharing both internally and with key external partners.
AI in Retail Demand Forecasting
One of Tolmare’s most promising AI initiatives involves a predictive algorithm designed to help retail outlets manage inventory more effectively. Traditionally, sales agents visited stores every few weeks, often discovering half-empty coolers. To address this, Coca-Cola developed an AI system that combines historical retail scan data with weather patterns and geolocation information from Google. This allows for more accurate predictions of future sales.
The AI insights are used to send targeted messages via WhatsApp to store managers, advising them when to restock specific products like Sprite or Diet Coke. The pilot program, tested in three countries, resulted in a 7% to 8% increase in sales compared to non-AI users. Tolmare plans to expand this tool to more markets globally, aiming to improve efficiency and reduce waste.
AI in Content Creation
Another major AI application at Coca-Cola is in content creation. With operations spanning over 180 countries and marketing materials in more than 130 languages, the process of creating culturally relevant content is both time-consuming and expensive. Recently, Coca-Cola generated 20 AI-created assets based on its proprietary intellectual property and produced 10,000 variations for use across different regions and languages.
According to Tolmare, consumers are 20% more likely to engage with AI-generated content compared to traditional methods, and the production time is three times faster. However, he stresses the importance of human oversight. “AI has proven that it can unlock value that a human being would not be able to unlock because the computation needed to mine through all this data and bring something meaningful is so complex,” he says.
Despite these successes, there have been challenges. A previous AI-generated Christmas ad faced backlash, and another campaign mistakenly included a quote from a book that author J.G. Ballard never wrote. Tolmare acknowledges the risks but notes that Coca-Cola has strict guidelines to prevent harmful social biases and to detect hallucinations or deepfakes in AI-generated content.
A Career in Technology
Tolmare has spent over two decades working in the technology industry, including roles at Palm, Cisco Systems, and HP. He joined Coca-Cola in 2018, shifting from developing technology to leveraging it for business growth. One of his early initiatives was a full-scale move to cloud computing, retiring or selling all of Coca-Cola’s physical data centers. Today, approximately 80% of the company’s infrastructure runs on Microsoft Azure, with the remainder split between Amazon Web Services and Google Cloud.
His approach to generative AI experimentation mirrors this cloud strategy. Coca-Cola collaborates with Microsoft and OpenAI, while also exploring other AI providers such as Google, Meta, and Anthropic. Tolmare believes in avoiding over-reliance on a single vendor, emphasizing the need for flexibility as the AI market continues to evolve.
The Future of Agentic AI
Looking ahead, Coca-Cola is exploring the potential of agentic AI—systems capable of performing tasks autonomously or with minimal human intervention. The company is evaluating offerings from vendors like Microsoft, SAP, and Adobe, while also developing its own AI agents trained on internal data. These systems remain in the pilot phase as Tolmare assesses their cost-effectiveness and ability to deliver optimal business outcomes.
“We haven’t launched agentic AI in production yet, but we are very close,” Tolmare says. “I’m fascinated by what this can do for our business.” For now, the focus remains on ensuring that AI investments align with Coca-Cola’s long-term goals of efficiency, scalability, and innovation.
