AI Adoption: Closing the Gap in Workforce Transformation

In Mexico, the question of whether companies will adopt artificial intelligence is no longer hypothetical. In May 2025, Microsoft reported that 89% of Mexican business leaders planned to integrate AI agents into their teams during that year. Forty-one percent were already using agents to automate workflows or business processes, and 65% of workers reported using AI as a workplace tool.

The speed of adoption is undeniable. Yet another question deserves equal, if not greater attention: What happens to people when technology arrives faster than an organization's capacity to absorb it? That is where a less visible side of digital transformation begins, and it is the side almost no one measures.

Implementing a tool does not mean a company has transformed the way it works. There can be licenses, users, pilot programs, and new features without a single meaningful decision changing. The risk is particularly acute with artificial intelligence. A company can make a copilot available to hundreds of employees and consider the implementation complete. But if those employees do not know when to use it, how to verify its outputs, what information they can share, or which decisions remain under human accountability, the technology is present, but the transformation has not yet begun.

The data reveals this gap, though it is worth noting where it comes from. A McKinsey study published in 2025, focused on the U.S. market, found that 48% of surveyed employees believed formal AI training would increase their day-to-day use of these tools. Yet only 22% said their organization provided anything beyond minimal or no support in developing AI-related capabilities. It is not Mexican data, but it describes a pattern any technology leader in our market would recognize instantly. Integration of these tools into daily workflows also ranked among the top factors that could drive greater usage.

The problem, then, is not simply training. It is about guiding change—something that demands far more than a course, a corporate webinar with cameras on, internal emails destined for oblivion, or a user manual.

New technology alters tasks, but it can also alter criteria, responsibilities, and execution timelines. If the organization does not explain what is changing and why, each person will construct their own interpretation. That is already happening. In the same study, McKinsey found that employees were three times more likely to use AI in more than 30% of their daily tasks than C-level leaders estimated. Thirteen percent of workers reported doing so, compared to an executive estimate of just 4%.

The gap is revealing. While leadership may believe adoption is still nascent, employees are already experimenting with these tools. And when that adoption occurs without clear direction, the company forfeits the opportunity to decide how it wants technology to enter its operating model.

This divide is not confined to a single country or function; it also appears across sectors. The same report ranks transportation and logistics among the industries investing least in AI, alongside financial services and energy, a signal that should raise alarms given the transitions underway among our clients and, of course, the constant evolution of the end user. The pattern repeats itself: it is not that the tool does not exist; it is that the decision of what to use it for and how to measure it has not yet been made.

And that pending decision carries a measurable cost, even if it rarely appears in the implementation report. It is the difference between installing a system and redesigning a process.

In logistics, that difference is visible every day. Automating fleet routing is not the same as deciding what information that system will use to prioritize a delivery, who reviews an exception when the algorithm gets it wrong, or what happens to the dispatcher's judgment that previously drove that decision. The tool can be fully operational, and yet no one in the organization may yet know what to do with what the tool is telling them.

That said, a technological transformation cannot remain confined to the IT department. Technology can select, integrate, and deploy a tool, but it falls to leadership to define what problem it wants to solve, what outcome it expects, and what must change across the organization to achieve it.

That requires concrete decisions. If a company adopts AI to improve productivity, it must determine which productivity it wants to improve and how it will measure it. If it seeks to accelerate decision-making, it must identify which decisions and what information they require. If it aims to free up time for its teams, it must decide what people will do with that time.

We must keep in mind that technology will continue advancing at a pace that will rarely wait for organizations to feel fully prepared. But being prepared does not mean knowing every tool. It means having clarity about the problem you are trying to solve, engaging the people who will have to change the way they work, and building the capabilities required for that technology to deliver real value.