Until a few years ago, software developers were the clear winners of the technological revolution. Within a decade, jobs in the field more than doubled, reaching 1,895,500, according to the U.S. Bureau of Labor Statistics. The agency estimates that this number will continue to grow through 2034, outpacing other professions, but at a more moderate rate (15%). The idea that anyone pursuing this path would have a guaranteed future has become obsolete.
Today, the real differentiator is no longer access to technology, but the ability to execute with it. We are now facing a new paradigm.
AI’s real disruption is in how little of that potential is actually being realized.
A recent report by Anthropic on the future of the labor market offers a useful starting point for understanding where we are heading. Based on the usage of Claude and the potential of language models across specific activities, the study estimates which occupations are most exposed to automation, using data from O*NET. However, beyond exposure, the real gap lies between potential and actual value creation. In all cases, the creation of real value remains significantly below what could potentially be achieved.
Among the most exposed occupations are software developers, customer service representatives, medical records specialists, market and investment analysts, and data entry operators. A key insight is that AI’s potential is measured in terms of specific tasks rather than the outright replacement of human jobs. The challenge — for both companies and individuals — is to understand this shift and adapt in ways that allow them to remain relevant across the value chain.
The Evolution of Roles and the Limits of Access
For developers, the report specifically highlights tasks such as writing, updating, and maintaining software programs. However, this does not imply that the profession is at risk of extinction. On the contrary, one of the most valuable assets in this new paradigm is technological literacy across all areas of an organization, making a transition toward new roles both necessary and natural. The ability to apply soft skills (or power skills), referring to interpersonal and behavioral abilities such as communication, collaboration and emotional intelligence will be essential for adaptation.
Never before in history has technology been as accessible as it is today. It is available to everyone at a low cost, with transformative potential ranging from building a basic website in minutes to discovering new drugs for medical treatments. Redefining processes is far easier than it was five, ten, or fifty years ago. But accessibility alone does not guarantee impact.
The Execution Gap
This leads to the second pillar of the new paradigm: implementing AI is not enough. The 2025 “State of AI in Business” study by MIT already warned that 95% of AI pilots fail, largely due to how they are adopted. Simply acquiring a license and leaving its use to individuals is insufficient, instead the challenge is not access to AI, but its integration into core processes. The Anthropic report, with its focus on tasks, provides a useful framework for identifying where to focus and how to design strategic solutions by effectively leveraging technology.
From Experimentation to Real Impact
When Julio César crossed the Rubicon, he knew there was no turning back. He declared that the die was cast and carried out his mission. Many companies now find themselves at a similar point with regard to AI: they must design and implement a strategy to improve efficiency, or risk being overtaken by others who will. There is no turning back.
The question, then, is how to move forward. The first step in this stage of AI agents is to define measurable ROI. Identifying the point of friction and quantifying the impact of what is implemented marks the difference between real solutions and “tech washing.” Data must be cleaned and integrated.
In this process, talent is essential to orchestrate the new systems and guide them toward autonomy. Only once this first phase is resolved can implementation scale, while human oversight continues to ensure that the AI agent’s learning and evolution align with what is needed.
The healthcare industry provides one of the most compelling examples of how this new paradigm can literally change lives. The market is valued at approximately $1.5 trillion, with expectations of continued growth in the coming years, partly driven by AI. A recent example is PharmaMar, a leader in the discovery, development, and commercialization of drugs, which is implementing AI agents to reduce costs and timelines in the development of new molecules and the execution of clinical trials.
This is not just about adopting AI, but about embedding it into critical workflows to accelerate results. It would be inaccurate to attribute this solely to AI; rather, it reflects the strategic vision of a company that leverages available technologies to analyze vast volumes of data—data that have been studied for decades by thousands of scientists—in order to accelerate the treatment of patients.
This new paradigm demands adaptation. The die is cast. Success will not come from merely adopting technology, but from executing it in a way that generates real value. This means moving beyond experimentation and focusing on how AI impacts productivity, efficiency, and decision-making in real-world scenarios.
Failure will come from a lack of understanding this. Those who internalize this shift early will define the standards of the next era, the rest of the people will be in the other side of the Rubicon.
