The Gut Check Moment for AI

When you have imperfect information, perfection is your enemy
compass hand

By Elspeth Murray

 

Artificial intelligence is a fundamental enabler. It will drive most organizations into step-function or tectonic changes to everything they do, both in their core business and in new businesses they create. We see this fundamental shift in where money is to be made and in the speed of change. But existing processes in many organizations are not keeping pace — lines of approval and committees introduce lag — and confusion abounds. Does anyone know what AI will look like in five years or even next year? Therein lies the challenge in making decisions and placing bets. No one knows.

So how do you make decisions in a world of uncertainty? If you’re Google or Facebook or any similar firm, your response would be to fund plenty of experiments, produce scenarios, or perform probabilistic modeling. You would pay attention to which ideas were working and which ones were not. Those that were promising and linked to an identifiable business imperative would get more resources.

In the automotive industry, for example, firms in the coming years will earn more revenues and profits from things that don’t actually exist today, thanks in part to AI. But there are many versions of what the future can look like. Executives at Ford and GM are managing this transition by placing  bets that, in the near term, most of the value to be extracted from AI will come from within their core business and existing processes: post-purchase cross- and up-selling, predictive service recommendations, in-line quality control, virtual prototyping and testing. The really big moves in terms of autonomous driving mobility services will happen later. They are testing and investing accordingly.

In many organizations, just getting the licence to take a risk with imperfect information is difficult

This approach can work well. But in many organizations, just getting the licence to take a risk and move forward with imperfect information is difficult. The leaders in these organizations are more comfortable commissioning a bunch of reports that would purport to reveal the future rather than turning around the question to ask what they would have to believe in order to say yes and then trying to answer that question.

The dilemma in cutting through the AI uncertainty is a question of leadership. Leaders have to gather all of the information they can possibly have at their disposal and then actually make a call. The gut check moment is when they have to take complicated viable options and distill them into actions.

When you have imperfect information, perfection is your enemy. It is an understandable impulse in organizations that are very successful in highly regulated industries, but it is a massive barrier to investing in a timely fashion and just getting moving. Plant your seeds and learn what you must. Short list the best options. Set firm time limits on discussions that focus peoples’ attention before making a decision.

If all else fails, put a quote on the wall in your office that says, We will not undertake a quest for perfection because it will kill us. Otherwise, you might as well sit in your foxhole and wait for a competitor to take you out.

Learn about the Queen’s Master of Management in Artificial Intelligence

Smith School of Business
Goodes Hall, Queen's University
Kingston, Ontario
Canada K7L 3N6

Follow Smith Business Insight on: