Data structures: a strategy to govern

Information can be understood as a rearview mirror—a mechanism for looking back, organizing results, and understanding what has already happened. Or it can serve as a platform for decision-making: a new lens through which companies can envision the future.

For a long time, managing strategy meant setting objectives, allocating budgets, overseeing teams, and measuring results. But how do analytical models fit into the equation? What happens when some decisions are made by people while others are influenced by systems that prioritize variables, detect patterns, and recommend courses of action?

The answer may sound familiar: we need to pay close attention to the structures, or rails, through which data flows. It is not just about what information flows through the organization, but also how it flows, under what parameters, and across which areas.

The OECD Economic Outlook 2025 warns that growth and productivity, in both advanced and emerging economies, will increasingly depend on the ability to integrate digital technologies into production processes in a structural, rather than superficial, way.

When a system prioritizes certain suppliers, adjusts inventory levels, identifies financial risks, recommends an investment, or predicts fluctuations in demand, there is an underlying structure organizing those decisions.

Models are not neutral. They depend on how they are trained, which variables they prioritize, which data they consider relevant, and which data architecture they operate within. That logic has an impact on resource allocation and a company's competitive capabilities.

When decisions are made through models, the data structure becomes the instrument of organizational governance.

One of the most common misconceptions about artificial intelligence is to think about it solely in terms of technological efficiency. Yet its deeper impact emerges when it changes the structural variables of a business.

The World Economic Forum's AI at Work 2026 report argues that the next stage of AI adoption will be shaped not only by individual productivity, but also by the ability to transform entire organizational structures.

Organizations that establish clear rules around data quality, interoperability, traceability, and model oversight will be better positioned to scale artificial intelligence without sacrificing operational consistency.

An integrated data architecture makes it possible to optimize inventory, anticipate supply chain disruptions, reduce financial costs, improve demand planning, and make more accurate business decisions. Above all, however, it enables organizations to respond more quickly to volatile conditions.

This is particularly important in Latin America, where many companies operate in unstable macroeconomic environments characterized by currency fluctuations, inflation, and supply chains that are highly sensitive to external shocks.

The UNCTAD World Investment Report 2025 highlights digital infrastructure and the ability to adopt new technologies as increasingly important factors in attracting investment and maintaining competitiveness in emerging economies.

For a long time, a company's strategic assets were relatively easy to identify: financial capital, physical infrastructure, intellectual property, or market positioning.

Today, data is emerging as a new way of thinking about assets. But the key difference is that data does not create value on its own. Value emerges when there is a structure capable of connecting data to processes. And it is information flowing and interacting across the organization that creates the synergistic effect capable of generating value. AI echoes this logic, but it does not eliminate the need to build strong organizational systems. If anything, it makes that need more visible—and more urgent.

This is likely where one of the biggest challenges for Latin American companies lies, because digital transformation means redesigning how information flows throughout the organization.

Data structures are beginning to function as a new mechanism for strategic control.

They do not replace human leadership or executive decision-making, but they do redefine the framework within which those decisions are made. As models play an increasingly important role in resource allocation, governing strategy also means governing the criteria under which those models operate.

For years, companies focused on controlling processes. Today, they need to start controlling decision-making systems. This requires at least three shifts in perspective.

• From accumulating data to governing it. Not all data carries the same weight, nor should it influence a model in the same way. When an organization operates with fragmented, outdated, or inconsistent information, AI ultimately amplifies those shortcomings. A model working with incomplete information can optimize the wrong decisions at scale. That is why organizations need to establish clear criteria for which data enters their systems, who validates it, how it is updated, and which standards govern its use.
• Integrating models into real-world operations. In many organizations, there is still a gap between analytics and execution. Data teams produce insights, while operational teams continue to make decisions according to different logic or timelines. But when models are connected to financial, procurement, workforce management, or supply chain systems, they stop producing isolated recommendations and begin to influence the actual allocation of resources. That step requires interoperability: different systems need to communicate, share information, and operate according to a common organizational logic.
• Making model-driven decisions auditable. Controlling a strategy mediated by artificial intelligence does not mean manually reviewing every decision—something that is impossible at scale. But it is essential to understand the criteria behind a recommendation, which variables carried the most weight, and which assumptions underpin the system. If a model recommends reducing inventory, changing prices, or prioritizing certain investments, the organization needs to be able to trace the reasoning behind that recommendation.

The new challenge is to design the rules under which this intelligence operates. The OECD Economic Outlook 2025 links the next wave of productivity precisely to the deep integration of digital technologies into companies' real-world operations. Because when models begin to participate in business decisions, controlling strategy also means controlling the invisible rails that organize those decisions.