Setting a goal to adopt AI is counterproductive. Hear me out.
A tool in search of a problem
When we begin with “We need to implement AI”, we tend to look at the workflows we already understand. We find a shiny new way to do something that was working reasonably well.
Meanwhile, the stubborn problems—the expensive, awkward, perpetually deferred ones—remain untouched. They’re still sitting in the corner where we left them.
Don’t start with where AI fits. Start with the problems you’ve been unable to solve.
Look at the quagmires
What have we been unable to solve? Where do our teams spend energy compensating for broken processes? Which customer needs have we considered impossible to serve economically?
Those questions give a new capability a chance to create genuinely new value. AI may help. It may not. Either answer is more useful than a technology adoption scorecard.
Don’t forget your strengths
I also ask where competitors might use new capabilities to undermine something we currently do well. It’s not enough to shore up weaknesses if the ground beneath our strengths is moving.
This is the same product work it has always been: understand the situation, find the leverage, make a sensible bet, and learn from the result. AI changes the available options, not the fundamental job.