AI and Design · AI Builders · November 2026
Asked why companies can't simply cut 20% of staff through automation, Hutka says the usual result is over-automating and then rehiring, drawing on the classic "ironies of automation" research.
…products and therefore adopting and coming back to use your products. So maybe I can start with the first bucket of those things. Does that feel like a good place I can get into maybe why you shouldn't automate all the things? I can also start with the business side if that's more a better spot.
…automate things for our employees? Can I cut 20% of my workplace? Can I streamline people doing, you know, taking meetings or writing emails when they should be doing other kinds of work? Like, that to me feels like where business is oriented around automation. Yes. Am I wrong? Am I wrong in that?
I don't think you're wrong in that. I think there is something around the, I guess when you're saying, you know, let's cut 20% of our workforce, you're going to immediately see, you know, less costs because you have 20% less of your workforce there. And that's the theory, right? And that is maybe very appealing in the short term because it is something that's very tangible. But I think there's the question of what happens after that. And the common pattern, I talk about this in the book too, is you over-automate, you let go too many people, and then you need to start bringing people back. And why is that happening? And this maybe gets into is letting 20% of your workforce go, like, is that actually the best way to move forward as a company? And I think there's a lot of maybe hidden gotchas in doing that beyond obviously like the real important cost to employees and letting people go. But if you're about to make that decision, I think there's a couple of things around what happens when you try and automate without having humans in or on the loop. And I think this goes back to like going back to human…
I think there is a tendency to think of jobs as being one primary task. And perhaps it is actually multiple tasks wrapped into one, like being a designer, for instance. Like, well, the pushing of pixels on production part, well, that is a primary task, but it is certainly not the only task that I…
Yes. This was an unlock when I was writing the book about like I felt I really needed to introduce that concept early. And the more I got into the book, the more I was just reading and adding in economics literature into the mix. I've spoken a lot about the importance of adding economics to that…
What you're talking about there sounds like a really nice process to think about things and actually come up with what should we automate? What should we not? I think you discussed that in the book, right? And you discuss actually a model for folks to kind of think through this in a structured way.…