Efficiency Is Not a Strategy
AI is giving hours back. Almost nobody decided what they are for.
My thought of the week is, if AI is saving me all this time, where did the time go?!
The promise of AI at work, the one in every deck and every vendor demo, is time. Real hours back, and I think the hours are real. The part I keep coming back to is that hardly anyone can tell you what actually happened to them. They exist on paper, but they do not show up on anybody’s calendar.
This week I read Sahil Bloom’s Jevons Paradox: The Perils of Efficiency, and he gave me the words for what I’ve been thinking about. Jevons Paradox is an economics idea from the 1860s, originally about coal, and the short version is that when you make something more efficient, people consume more of it, not less. His example is the one that resonated. “Smartphones made email processing more efficient, so we all send 10x as many emails as we did before.” Then he takes it one step further, “get more efficient at processing email, and you’re effectively reducing the cost to others of emailing you, so it’s more likely they will.”
When I replace e-mail with AI in that sentence, it all comes together. Every hour you get back lowers the price other people pay to ask you for something, so they ask for more, usually without even thinking about it, because you made it cheap.
That is the individual version. The organizational version turned up in Fortune in Dan Schawbel’s How AI turned your best work into the bare minimum. It is blunter than I expected from a business magazine, and I’m here for it. “Efficiency gains from AI are not being returned to employees as breathing room. They are being immediately converted into higher output expectations.” The numbers he pulls together are wild. Employees are saving two or more hours a day with AI. 60% feel more pressure to be productive, not less. Focused deep work is down 9% and 77% say AI generated work takes more review time than human work, which tells me some slice of that two hours was never really saved at all, it just changed clothes and moved to a different part of the day.
The line that really stuck it to me was “the productivity gain and the human performance cost are arriving simultaneously. Most organizations are only tracking one of them.” That, my friends, is not an AI problem, that is a measurement problem, and we have been making that exact mistake through every transformation, reorg and system migration I have been part of in my years in corporate...this one just moves WAY faster.
Shifting gears a little.
Brené Brown did a solo episode of The Curiosity Shop on August 20 called Brené Brown on Protecting Our Attention, Time, and Humanity in a World That Won’t Slow Down, and she takes on the quality side of all this (for anyone new here, I am a HUGE Brené fan). She talks about where AI can support our work without replacing human discernment, and she names “work slop,” the polished output that costs a leader real cognitive bandwidth to untangle. The question she puts to leaders is a pain point I think we can all relate to. Can you tell the difference between polished output and actual understanding? She also tells a story about getting fired from a house cleaning job and what it taught her about why shiny can never replace mastery.
I think that might be the whole thing, honestly. AI is very good at shiny. Mastery is still ours, and checking for it takes time we just finished telling everybody they saved.
Next, the reason I do not think a policy fixes any of this. I referenced this article last week too, but I keep coming back to it around a variety of topics. Adam Grant wrote The Truth About People Pleasing in Granted, and he is not writing about AI at all, but he describes the mechanism exactly. “The root of chronic people pleasing is not concern for others. It’s concern for their acceptance and approval.” Hand a team two hours back and say nothing about what those hours are for, and the hours go to whoever asks first. It’s not because your team is weak, it’s because saying no costs something, and nobody wants to be the person who looks slow in the year everybody got fast. His answer is worth borrowing and saying out loud to your people. “Sharing your preferences is not a display of disrespect. It’s an expression of self-respect.”
Land the plane, Andrea...ok, so efficiency is not a strategy, it is a result, and somebody has to say out loud what the result is for, or the default answer is always more. If you have not said the words “the time we get back from this is for (fill in the blank)”, you have not given your team time back, you have given them capacity, and capacity gets filled.
I spend my working life helping organizations adopt this stuff, and I still catch myself admiring the hour I saved instead of asking what the hour was for. It is a very easy thing to miss when the tool works.
Pause and reflect
As we head into the weekend, take a few minutes with these.
- 01Where did your saved hours actually go this week? Not where you meant them to go, where they actually went.
- 02What is one sentence you could say to your team on Monday that finishes “the time we get back from this is for...”?
- 03What did you approve this week because it looked polished, and are you sure you understood it?
I’ll leave you with this -
“The single greatest challenge for any high-performing, ambitious person is eliminating the guilt associated with free time and rest.”
Have a great weekend!
Andrea