Throughout the history of the software industry, the logic that organized it followed a relatively stable pattern. Human development teams wrote code, implementation cycles tested and deployed it, and then the final product evolved through successive versions. That model is no longer at the center of the industry.

Artificial intelligence and generative models in particular—have introduced a disruption that goes far beyond task automation. AI accelerates processes such as software development and code generation, but it also redefines the very nature of software production. Software is no longer simply a product; it is becoming a dynamic capability that depends on data, models, and usage contexts.

From code to intelligent systems

The first shift is structural. Now that code can be partially generated by AI, the bottleneck has moved from writing code to designing systems that work effectively. In that sense, code is much like information: you can have a lot of it, but if it is not used correctly and within a context that can leverage that data, it serves no real purpose.

As a result, the focus has shifted toward data integration, rule definition, model oversight, and ensuring the quality and traceability of outcomes. This has led to a more systemic approach that extends far beyond the application itself.

Within this framework, data plays a foundational role. Software is no longer the programming of a static sequence of instructions; it is a dynamic system that learns and adapts based on the data it processes. The quality, availability, and representativeness of that data determine the type of intelligence that can be built. However, not every organization can deploy the same AI systems or capture the same value from them. As a result, the ability to integrate and govern data becomes a critical differentiator.

The Latin American Artificial Intelligence Index published by ECLAC shows that the region is making progress in AI adoption, yet significant gaps remain in infrastructure and governance. At the same time, estimates from the Linux Foundation suggest that up to 40% of working hours could be transformed by generative AI, primarily through the partial automation of tasks and the productivity gains that follow. The redesign of human work and data-driven processes is imminent.

The new software business model

As software becomes easier to generate, adapt, and scale, value is no longer concentrated in a closed product. Instead, it is shifting toward platforms, data, and intelligence capabilities. The true opportunity lies not in what AI gives us, but in what we can create together with these systems.

If code writing is no longer at the core, the skills required are changing as well. This does not mean starting from scratch. Traditional software development remains highly relevant, but it must now be complemented by new capabilities: systems design, data management, model oversight, and governance.

This represents a shift in mindset—from execution toward more critical, systemic, and relational thinking. Understanding how systems operate, how models behave, and how their outputs should be evaluated becomes more important than writing code from the ground up.

The transformation is not occurring uniformly. As noted in an analysis published by El Financiero, software development as we once knew it is entering a phase of profound change, where productivity gains are reshaping professional roles. Repetitive tasks are increasingly automated, while responsibilities related to design, oversight, and decision-making continue to grow.

This raises an important question: How do we transform talent in a context marked by skills gaps and uneven access to technology? The answer lies in strengthening uniquely human capabilities—the ability to combine technical expertise with an understanding of local context. It requires openness to hybridization and process customization. In practice, this means more interdisciplinary teams, more flexible processes, and closer collaboration between business and technology functions.

Redesigning how organizations work

Change never happens uniformly or overnight. Traditional models coexist with new ways of creating value. AI adoption is advancing rapidly across Latin America, often on top of structures that still reflect fragmentation and unequal capabilities.

As development becomes less centered on coding itself, software does not disappear. However, it loses its autonomy as a standalone unit of value and becomes embedded within broader structures.

In this context, the organization of work becomes the primary variable of adjustment. Highly structured and repetitive tasks will become increasingly automatable, while activities that combine context, judgment, and decision-making are more likely to evolve than disappear. What is emerging is a transition from teams organized around functions to more hybrid models that bring together technical, operational, and business profiles to manage intelligent systems. Will new forms of coordination emerge? Most likely. And if they do, it will be because the way we understand value—both human and technological value—has fundamentally changed.

Blue Diamond International, through RISE with SAP, built an environment of systemic and relational intelligence by creating a platform that processes information within a business context, unifies it, and connects it across functions. The company recognized the importance of AI as a proactive tool for improving revenue growth, asset management, and unified commerce.

A similar transformation took place at Volkswagen Mexico. Using SAP Datasphere and SAP Analytics Cloud, the company implemented a new financial information model that transformed previously fragmented methods and processes into an integrated system powered by data from multiple sources. In both cases, competitive advantage no longer comes from isolated development efforts. It comes from the ability to coordinate data, processes, and decisions across the organization.

Over the coming years, this distinction will become increasingly decisive, perhaps in ways we cannot yet fully measure. Organizations that redesign how they work—not just the technologies they use—will capture sustained productivity gains. Those that do not will simply layer intelligence onto structures that remain unchanged.

That, in my viewpoint, is the optimistic future of software. It is no longer the center of the business, but an essential capability that enables the business to operate. Software's role is evolving into the foundation through which an organization's data is generated, processed, and circulated.

How will we allow these systems to shape our decisions, processes, and data flows? How will we think and generate knowledge alongside them? Perhaps that is where the next divide will emerge: between those who learn to operate with AI and those who do not.