A Fitbit, a broken air conditioner, and the real test of your AI strategy
What AI makes cheaper may not be what matters
In David Sedaris’s essay Stepping Out, he describes what happened when he got a Fitbit.
It started reasonably - ten thousand steps a day, a gentle tingle on his wrist when he hit the target. Then fifteen thousand. Then twenty-five thousand. By the end he’s doing sixty thousand steps a day, picking up litter along the roads of Sussex, and when his Fitbit dies he lasts five hours before ordering a replacement, hands shaking as he opens the box.
Sedaris started walking because he enjoyed it. Then the measurement replaced the enjoyment as the point.
He stopped asking “is this a good walk?” and started asking “is this enough steps?” The answer was never yes.
There’s a story in Brad Jacobs’s book that runs in the other direction.
Jacobs has built multiple billion-dollar-plus companies from scratch, but the best bit in his book is about something considerably more mundane.
His home AC broke during a heatwave. He called Steve, his repair guy, who came on a Sunday without hesitation.
Steve had originally worked for the company that installed the system - a large outfit that got bought by a private equity firm - not one of the good ones - which cut costs, stripped the service back, and resold the business.
Steve saw what was happening and left to set up on his own.
“How’s the old company doing?” Jacobs asked him.
“Terrible,” Steve said. “There’s no pride in their service. They’re giving the industry a bad name around here.”
Steve isn’t perfect. He’s misdiagnosed problems. He’s ordered the wrong part. But he owns up to it immediately and fixes it fast. Customers flock to him. The company he left is haemorrhaging clients.
The PE firm did what Sedaris did with his Fitbit.
They found the thing they could measure - cost - and optimised for it until they forgot what the business was actually there to do. The service got cheaper to run. It also got worse. The people who cared about doing good work left. The customers followed them.
The cost case
The same pattern is playing out in many AI conversations right now.
Cost reduction is the easiest business case to make.
“We’ll save X by automating Y” sounds like a no brainer.
“We’ll make our best people more effective” is harder to put a number on. And much harder to actually achieve.
So the cost case wins, and six months later you’ve got a cheaper operation that nobody - customers or employees - feels good about.
Microsoft’s latest Work Trend Index, based on 20,000 AI-using workers across ten countries, backs this up. The organisations getting value from AI aren’t just handing people tools or using them to reduce headcount. They’re changing the culture, management habits, and workflows around them.
Steve’s old company had the same technology available as Steve. The difference was everything around it.
The better question
So, the question to ask yourself is really quite simple:
When the AI’s been there a while, are your best people more valuable or less?
If the technology handles the routine so your best people can focus on the stuff that only they can do, you’re building Steve’s business.
If the people with experience and skill matter less than they used to, you’re building the bad version of the PE story.
And your best people will notice before your customers do. Steve didn’t leave because the PE firm fired him. He left because they made it impossible for him to do his job the way he knew it should be done. The people you most want to keep will do exactly the same thing.
Sedaris, at least, only ruined his own walks.
Thanks for reading.
Ollie
Ollie on Work is a weekly newsletter about what I’m learning from building with AI, advising leadership teams, and trying to bridge the gap between what technology can do and how businesses actually work. If someone forwarded this to you, you can subscribe here: