Rise of the Strategy Machines

Editor’s Note: This article is one of a special series of 14 commissioned essays MIT Sloan Management Review is publishing to celebrate the launch of our new Frontiers initiative. Each essay gives the author’s response to this question:

“Within the next five years, how will technology change the practice of management in a way we have not yet witnessed?”

As a society, we are becoming increasingly comfortable with the idea that machines can make decisions and take actions on their own. We already have semi-autonomous vehicles, high-performing manufacturing robots, and automated decision making in insurance underwriting and bank credit. We have machines that can beat humans at virtually any game that can be programmed. Intelligent systems can recommend cancer cures and diabetes treatments. “Robotic process automation” can perform a wide variety of digital tasks.

What we don’t have yet, however, are machines for producing strategy. We still believe that humans are uniquely capable of making “big swing” strategic decisions. For example, we wouldn’t ask a computer to put together a new “mobility strategy” for a car company based on such trends as a decreased interest in driving among teens, the rise of ride-on-demand services like Uber and Lyft, and the likelihood of self-driving cars at some point in the future. We assume that the defined capabilities of algorithms are no match for the uncertainties, high-level issues, and problems that strategy often serves up.

We may be ahead of smart machines in our ability to strategize right now, but we shouldn’t be complacent about our human dominance. First, it’s not as if we humans are really that great at setting strategy. Many M&A deals don’t deliver value, new products routinely fail in the marketplace, companies expand unsuccessfully into new regions and countries, and myriad other strategic decisions don’t pan out.

Second, although it’s unlikely that a single system will be able to handle all strategic decisions, the narrow intelligence that computers display today is already sufficient to handle specific strategic problems. IBM Corp., for example, has begun to use an algorithm rather than just human judgment to evaluate potential acquisition targets. Netflix Inc. uses predictive analytics to help decide what TV programs to produce. Algorithms have long been used to identify specific sites for retail stores, and could probably be used to identify regions for expansion as well.