Your AI Doesn’t Need More Data, It Needs a Strategy

Why isn’t artificial intelligence living up to the expectations so many companies, investors, and tech leaders have been promising? For those of us working in the AI field, the applications are clear and continue to grow. But for leaders outside this space —even though the desire to use AI is there— the results they expected to see by this point in 2025 simply haven’t materialized.

Close to 90% of large companies use AI —especially generative AI— across their operations. So the inevitable question is: what’s going on? From what I’ve seen in organizations across more than a dozen countries, everything points to the same underlying issue that shows up across the entire organization and its AI goals: an AI strategy that is unclear, incoherent, or simply nonexistent. Let me explain.

AI is a tool, much like a saw, a hammer, or a wrench. Imagine I give you the following instructions to put your tools to use: “Build a 300-square-meter house with a budget of 45 million pesos.” At first, the project sounds simple. But then you realize you’re missing crucial information. You don’t know what style the house should be, who it’s for, where it will be built, or what the occupants prefer. And even more importantly, each of the millions of Fast Company México readers would likely give completely different answers to each of those questions. That would lead to entirely different home designs.

You might also think: “With 45 million pesos, I’d build a 600-square-meter home!” And you might be right; it would depend on where you build it. Sometimes the budget is more than enough, sometimes it’s not —and that’s exactly the point. How would you know? Your strategy would tell you.

Lots of Noise, Not Much AI Strategy

The absence of a coherent AI strategy is even more costly than in the house example. Many companies (though few would admit it) believe that spending on AI for a variety of isolated tasks is a viable way to “win with AI.” The result? Everyone across the company ends up building something different with the same tools, and leaders are left wondering what their teams are actually doing.

Between poorly executed cognitive offloading, drops in productivity, internal confusion, and doubts about the real usefulness of AI, problems pile up at both the executive and operational levels. This explains, in large part, why 95% of generative AI projects fail. There’s a lot of movement, but not much progress.

That’s why the beginning of any AI initiative requires a thoughtful, well-designed strategy —one that gives you the clarity to select the right project, assemble the right team, choose the right tools, size the investment properly, and train your people to use AI strategically. So what can be done? Is there hope?

Fortunately, yes.

From Chaos to a Plan

Print this out and sit down with your team to answer the following questions as a starting point for building an AI strategy. It won’t be a full strategy, but it will give you the most important calibrations to identify the right direction and the optimal methods for building an exceptional AI future.

First: What kind of company will your organization need to become to thrive in an era defined by AI competition and AI-enabled firms? For instance, what will your main competitors and the startups in your industry likely do with AI? And what would that mean for your business?

Second: Is your company one that uses AI from third-party vendors (the most common scenario) or one that aims to build proprietary AI systems (less common, but transformative)?

Third: Make two lists. One identifying the areas where AI could have the greatest impact in the near term, and another listing the areas where implementation will be difficult or time-consuming. Use the first to find your “quick wins” and the second to design your medium- and long-term strategy.

Fourth: If you read my previous article on luck (of course you did), you know how essential it is to understand which variables you can control, which you can influence, and which are completely outside your hands. Remember: all strategy contains elements of uncertainty and chance built into it.

If you apply the first three steps, you’ll be well on your way to building a stronger AI strategy than most companies —and you’ll be saving a meaningful amount of resources. And if you fold in the lessons from the fourth step, you’ll be far ahead of your industry, because you’ll understand how to think about your strategy at a deeper level.

AI can be incredibly powerful and incredibly useful… when used with clear strategic intent. I always tell my team, clients, and students the same thing: fall in love with the problem and with the user at the center of the AI system. Understanding what you’re building and who you’re building it for before you begin is, without question, the best strategy of all.

Originally published in Spanish for Fast Company Mexico:
https://fastcompany.mx/2025/10/10/ia-estrategia-datos/