AI Strategy

•

Tim Hillegonds

Letting Go of What You Just Learned

Your team needs to learn AI quickly and stay willing to question what it has learned. A recently mastered process should change when there is evidence of a better way to serve the customer.

A couple of months ago, I listened to an interview with product leader Peter Sellis and came away thinking about something he said that's stayed with me: how quickly can you let go of something you just learned?

In the AI world, we're constantly learning new skills. We get used to using them, and then a new capability changes how the work can be done—and, suddenly, some of what we just learned needs to be reconsidered.

Ironically, one of the superpowers of the AI era is being able to unlearn as quickly as you learn. However, that's a difficult tension for businesses to deal with. They need people to learn quickly, but they also need them to remain willing to question what they've learned and, in some cases, unlearn it completely.

That can be especially difficult if you've invested in training, established a new process, or told everyone that this is now how work should be done.

What Are You Actually Learning?

I think the right way to think about this is to understand what you're actually learning. Part of what people learn when they use AI is how to work around its limitations. They break a task into smaller steps, move information between tools, or spend time preparing materials the system can't yet work with directly. That can all be helpful, but it might not all be permanent.

For example, think about a team that develops a multistep AI workflow for customer proposals. Suppose a newer capability can combine several of those steps while meeting the same quality requirements. The original process still works, but continuing to follow it may no longer be the most useful thing for that team to do.

BCG describes a related difficulty: unfamiliar AI workflows can initially reduce performance, and unchanged expectations can push people back toward familiar routines. With AI, that familiar routine might be the one you only recently encouraged them to adopt.

This is why I think learning in the age of AI can't just be about mastering a particular process. It also has to help people think about the work itself. That means understanding why a process exists, what problem it solves for the customer, what a good result looks like, and who is accountable for its accuracy. Those things still matter when the sequence of steps changes.

I'm not saying everyone should start over whenever a new model comes out. What I am saying is that you should revisit things when there's evidence that the work can be done better, even if you just created the method. The real question is whether what you're doing still serves you and your customers, or whether you're continuing simply because it's what you've learned to do.

That’s the tension businesses have to manage: people need to learn quickly, put what they learn to work, and remain willing to reconsider it. Having only just mastered a process can’t be the reason to keep it when there’s evidence of a better way.

Sometimes, getting better at the work means letting go of the way you were just taught to do it.

Collage of three people holding trophies, with black bars across their eyes, binary digits, and lime-green and orange rectangles.

Related Insights

Stay Ahead With Powerful Insights

Get exclusive insights, actionable strategies, and ideas delivered straight to your inbox.

Stay Ahead With Powerful Insights

Get exclusive insights, actionable strategies, and ideas delivered straight to your inbox.

Stay Ahead With Powerful Insights

Get exclusive insights, actionable strategies, and ideas delivered straight to your inbox.

Have Questions?

Email us at info@six06strategy.com or submit an inquiry form.

Six06

Have Questions?

Email us at info@six06strategy.com or submit an inquiry form.

Six06

Have Questions?

Email us at info@six06strategy.com or submit an inquiry form.

Have Questions?

Email us at info@six06strategy.com or submit an inquiry form.

Six06