Juan Santiago | CEO at Santex and Technology with Purpose | Cofounder of Incutex | Partner at Kalei Ventures.
A CEO I respect recently told me his company had just signed a significant contract with one of the leading global AI platform providers. The benefit, of course, is faster deployments, better integrations and real productivity gains within weeks. I asked him one question: Where does your data go when the contract ends? He didn’t have an answer, and when I asked whether anyone on his team might know, he admitted nobody did.
That conversation is happening in boardrooms across every industry right now. Companies are adopting AI tools with genuine enthusiasm, yet with almost no clarity about who keeps the valuable information they generate in the process. Most treat it as a procurement detail rather than a governance one. I’ve come to see it as a strategic blind spot, and the window to close it narrows with every contract renewal. Organizations that don’t govern how their AI systems are trained, updated and eventually retired take on a slow, compounding risk in which the vendor holding the data and the model used ends up increasingly carrying more of the judgment with each renewal, until the terms are effectively owned by someone else.
The world has settled into three distinct postures toward AI. One region is building a rights-based framework, not to protect governments, but to shield citizens and organizations from the unchecked accumulation of data by entities accountable to almost no one they extract it from. Another has chosen maximum velocity over protection, with deregulation that serves interests well beyond its technology sector. The reason runs deeper than policy preference. AI is the first technology that can reach billions of people while speaking to each of them individually, understanding their patterns and assumptions closely enough to reinforce what they already believe.
The Edelman Trust Institute’s 2026 Trust Barometer for Latin America shows what this outcome produces at scale. In the region, three out of four people report what researchers call an insular mindset. Seventy-four percent say they doubt or refuse to trust people who think, live or get information differently than they do. Organizations struggle to make collective decisions when their people no longer share the same version of reality. AI tools that mostly confirm what a team already believes, rather than test it, don’t save time on the work that matters. They just make the same assumptions faster.
Latin America alone represents an estimated 667 million people, according to United Nations projections. That population hasn’t been invited to help design the terms of this negotiation, and the region hasn’t outlined a clear position. It has become the place where new AI models get tested in markets with lighter protection, where regulatory gaps shelter practices banned elsewhere, and where data gets extracted at scale from people who accepted terms they never helped set. A regional strategy would require someone to actually author it. So far, mostly, nobody has.
Most executives I talk to treat this geopolitical map as somebody else’s problem. In 27 years of building Santex, I’ve learned that assumption is usually the most costly mistake in the room. A few years ago we started asking every AI vendor a question that seemed obvious in hindsight but wasn’t at the time. What happens to our clients’ data and our own models the day we stop paying? Two vendors couldn’t answer clearly. We walked away from both, even though their tools were strong, because a platform that can’t explain what happens to your data after the relationship ends isn’t really answering the pricing question. It’s avoiding the ownership one.
Leading a company today takes deep industry expertise, real fluency in what AI can and cannot do and a working understanding of the geopolitical forces shaping who controls the infrastructure you depend on. Every CEO who wants to continue making their own decisions in 10 years must have all three. Most of the people I encounter today have only the first.
What surprised me most about that decision inside our own company wasn’t the vendors’ answer. It was how many people on our own team assumed the question didn’t matter, as long as the tool worked well and shipped fast. Getting past that took longer than picking a new vendor did. It meant making data ownership part of how we evaluate any AI tool, the same way we already evaluate security or cost, and holding that line even when the slower answer cost us a deadline.
If someone asked you today which parts of your organization’s intelligence you’ve already handed to a vendor, would you know the answer without checking?
Every contract you sign, and every platform you integrate, is already shaping who controls AI, who gets the value and whose business becomes the raw material for someone else’s strategy. Your company has a stake in that negotiation. Whether you help shape the terms comes down to the questions you ask before you sign.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?