Why More AI Use Cases Can Produce Less AI Value
The fastest way to waste money on AI is to make “using AI” the goal.
I understand why companies are doing it. The pressure is real, and it is coming from almost every direction at once. Boards want to know the AI strategy. CEOs want visible progress. Employees are experimenting on their own. Vendors have suddenly discovered that nearly every software feature can be described as “AI-powered,” sometimes with a straight face. So companies respond by looking for places to use it: sales, engineering, maintenance, quality, supply chain, customer service, cybersecurity, finance, HR. The list grows almost immediately.
The problem is that the number of possible AI applications is expanding much faster than most organizations’ ability to implement them well. That turns what looks like a technology opportunity into a very traditional strategy problem: there are far more things we could do than things we can realistically fund, staff, govern, integrate, and sustain.
When Every Function Wants an AI Initiative
This is where things start to get messy. One business unit finds a compelling use case. Another sees something similar and wants its own version. A functional leader realizes their roadmap probably needs an AI component. Before long, the company has twenty pilots, several platforms, multiple vendors, and a growing collection of PowerPoint slides describing future value.
Individually, many of those projects make complete sense. Collectively, they may be impossible. The same data engineers are supporting several projects. The same cybersecurity people are reviewing all of them. The same operating leaders are expected to sponsor multiple transformations while still running the business they already have. Funding gets divided into smaller pieces, executive attention gets spread across more meetings, and the people with the deepest understanding of the business become bottlenecks because every project needs them.
Nothing is technically ignored. Almost nothing receives enough concentration to become important. I see this happen outside AI too. I have a drawer at home filled with old charging cables because every single cable has a perfectly rational reason for remaining there. Surely, someday, civilization will depend on my ability to locate a Mini-USB cable from 2009. So I keep all of them. Eventually the drawer barely closes. Organizations are remarkably similar with strategic initiatives: every project can defend its existence, which is how you end up keeping all of them.
Ask a Harder Question
The natural question for leaders right now is, “Where can we use AI?” I increasingly think that question is too easy. Given enough time, the answer will probably be almost everywhere. A more useful question is: Where would AI materially change the economics, performance, or competitive position of the business?
That produces a very different conversation. Saving employees a few minutes writing meeting notes may be useful. Reducing unplanned downtime by 20%, improving yield on a high-cost production process, shortening engineering cycles, reducing inventory without hurting service levels, or materially changing how quickly a company responds to customers could be transformational. All are AI use cases, but they should not automatically receive the same strategic weight. Leaders therefore need to look beyond whether AI can perform a task. They need to understand how often the decision occurs, what the underlying economics are, whether the necessary data is accessible and trustworthy, whether the recommendation can actually change a workflow, and whether improvement would create meaningful differentiation. A technically impressive model that sits beside the actual operating process is still just an interesting model.
AI Makes Focus More Valuable
The uncomfortable part comes next: some good ideas have to wait. That can feel irresponsible during a period of rapid technological change because restraint easily gets mistaken for hesitation. Nobody wants to discover three years from now that a competitor moved aggressively while they were still forming a governance committee.
But trying to move everywhere at once creates its own kind of slowness. Capital is limited. Technical talent is limited. Clean, contextualized data is limited. The capacity of an organization to absorb new processes is very limited. Executive attention may be the scarcest resource of all, even if we rarely put it into the business case. That is why I suspect the companies that benefit most from AI will not necessarily be the companies with the most AI projects. They may be the ones willing to make sharper choices about where AI deserves disproportionate investment.
That means identifying a small number of areas where the economics are significant, the business problem is real, the conditions for adoption exist, and the organization is willing to concentrate talent, capital, data, and leadership attention. Everything else does not disappear. It simply moves down the list. AI is giving companies an extraordinary number of new possibilities. That is exciting, but possibility itself has never been strategy. The strategic advantage comes from deciding which few possibilities matter enough to pursue with real force, and then having the discipline to leave the rest alone for now.
