How Heineken's CIO Built an AI Strategy Around Operationalized Data

The wave of generative AI that followed ChatGPT's arrival forced technology leaders across industries to determine which applications would genuinely benefit their organizations. Frank Iannella, U.S. CIO and SVP for digital and technology at Heineken, has navigated this challenge by spending the past two years refining his beverage company's approach to artificial intelligence. The company's technology group was already exploring ways to convert data and insights into concrete business actions when generative AI entered the picture and became an additional tool in their arsenal.

Speaking at the Gartner IT Symposium/Xpo in October, Iannella explained: "Part of this is being enabled by this new technology — AI. It's a new part of our toolkit." The company has integrated AI systems across multiple business functions, including supply chain management and sales operations. Iannella and his team have restructured critical operational areas—particularly reporting, insights generation, and the breakdown of data silos—in ways that other technology leaders can adapt for their own AI deployments.

The framework

Long before AI became mainstream in corporate environments, Heineken was already working to convert data into actionable decisions. The process begins with identifying a signal—a specific operational challenge that can be addressed, such as detecting when inventory levels for a product drop too low. The second phase involves pinpointing the lever, or the action the business would execute in response. In this scenario, accelerating product delivery represents one possible lever.

Previously, Heineken performed this work manually, but AI enabled the team to recognize signals and levers more rapidly and implement corresponding actions. The next stage involved embedding these capabilities directly into their transactional systems to operationalize the findings. Following implementation came an evaluation phase in which Iannella's team measured whether automating signals and levers actually reduced stockouts, then refined the process based on the results.

Iannella reflected on the timeline: "It took us over two years to kind of figure this out, leveraging what we want to use as well as a kind of new AI stack to actually accelerate and make this process successful. It's the building blocks of decision intelligence." A crucial component of their technical infrastructure is an ontology layer—a structured, machine-readable framework that gives AI systems the connection points they need to function effectively. According to Iannella, this layer is where the most significant work happens: "If you want your AI tool to actually make the right decisions, that ontology layer is where you have to have most of your work. That's where you're actually configuring a lot of the settings to make it work."

Decision-ready data

While most CIOs understand the necessity of high-quality data for AI systems, Iannella's team discovered they needed to advance beyond that baseline. The company shifted its focus toward what Iannella calls decision-ready data—information specifically structured to align with the business outcomes the organization wants to achieve. For Heineken, this involves evaluating whether the data is current enough to detect signals like low inventory, whether it connects to the operational levers such as ordering and delivery systems, and whether the data quality is sufficient to enable action.

Measuring AI's business value often proves difficult, yet this dimension offers CIOs a concrete way to demonstrate return on investment. Iannella posed the key questions: "Is it accurate enough? Is it specific and timely enough? Can you actually link it to an outcome so you can actually start to measure it?"

Democratizing AI

Rolling out AI across an organization requires making the technology accessible to employees with varying levels of technical expertise. Iannella emphasized the importance of enabling workers closest to the actual work while maintaining appropriate guardrails. He described this approach as creating "safe places, and places where you may need to be a little bit more concerned." Heineken deployed Microsoft Copilot to its workforce and established a user community where employees share implementation examples across different departments.

The company's approach to democratized AI balances experimentation in certain functions with governance structures where caution is warranted. Areas requiring heightened oversight include sensitive data handling, external-facing applications, automated decision-making, and anything affecting customers. Iannella told CIO Dive: "There's still trust to be earned. AI systems are providing us with an amazing scale, they're providing the recommendations and actions, but we're not there yet. There's still certain places you need a human to validate."

Heineken has discovered its AI value by maintaining equilibrium between developing use cases tailored to its specific requirements, acting on the insights AI produces, and exercising prudence in high-stakes decisions. Iannella summarized his philosophy: "I think the way we started thinking about it was: Here's a new reality. You can fight it, or you can try to guide it."