We Automated Support and Lost Our Escalation Instinct

**Stacie Baird** (0:07)
Stop me if you heard this sentence in the boardroom last year, because I haven't stopped thinking about it. AI can do 80 percent of what X team was doing.
Hey humans, welcome to The HX Podcast. This is Stacie Baird, and this is the podcast around where we talk about the intersection of humans and technology, and how we can, I don't know, get in our get along shirt and do all this together. So this statistic and this episode is really going to be about a lot of data and evidence that we're seeing around the layoffs that are happening right now, things that looked right on paper. But what actually happens sometimes and what we're hearing in the market today. So in that first opening, I heard that statistic batted around a lot like 80 percent, 90 percent, 70 percent, right? Everybody nodded in agreement. The math seemed like it worked out. The headcount came out of the budget and on paper, on the paper, it looked like a win. But six months later, that same company was on a call and we were trying to figure out why decisions were taking longer. Not shorter, why things were slipping through. Just slipping, but slipping off the plate.
That's what we're talking about today. Welcome to a new series. I'm calling the AI headcount trap.
Let's name it to tame it. Naming it plainly because I think a lot of leaders are living in this right now, and we don't really have the language for this yet. Somewhere in the last two years, you probably made a call, maybe more than once. I get it. We're all in this together, where you looked at a role, looked at what AI could do now, and did the math and said, hey, AI can do it. We don't need the person. Look, sometimes that's totally right. I'm not here to tell you AI adoption was a mistake. I use it myself. It certainly wasn't a mistake. But here's the piece that gets skipped almost every single time in those conversations. The role wasn't just doing tasks. It was exercising discernment, making calls, small ones, invisible ones. This matters more than that one, for example, that discernment. This client needs a human touch. This number looks off. Go check it. That's judgment. And judgment doesn't show up on a task list. So when you cut a role, because the task was automated, you don't just lose the doer of the task. You lose whoever was standing there catching all the stuff that wasn't listed at all in the job description. Because over the last 20 plus years of my career, we all do more than is on our job description.
So that's the trap. And not that AI failed, that we forgot to do the front-end homework of this assignment. We forgot to ask, what else walks out the door when we cut this role? I want to ground this in data like we always do here at The HX Podcast because it's not a hunch. This is documented. There's patterning happening. It's a big one. OrgView surveyed business leaders and found that 39% had made employees redundant because of AI deployment. Here's a part that should stop you and make you think. Of that group, 55% later admitted the redundancy decision was wrong. Forrester landed on almost the exact same number, separate research body, separate study. 55% of employers now say they regret laying off workers due to cuts related to AI implementation. Robert Half, really one of the biggest staffing agencies in the world, surveyed nearly 2,000 hiring managers in the US and found that 32% had eliminated a role primarily because of AI and had already rehired for the same or similar position. CareerMinds found something even sharper. This study, over half of companies that made AI-driven layoffs had rehired within six months. So here's maybe the most telling data point of all from Harvard Business Review's survey of more than a thousand executives. Most AI-driven layoffs were based on what AI might be able to do someday. Not what it had actually demonstrated it could do. Imagine that we hired people and we said, I think this person's going to be able to do this someday. Let's hire them and wait and see what happens.
Over 600 executives admitted on record. They had cut staff for anticipated future capability, not performance, not current performance, not past performance, but potential future capability. This isn't a fringe situation that only a few companies made this mistake. This is a majority pattern of how AI linked headcount decisions got made across industries, across verticals, big and small, top to bottom. Bet first, verify after, later or after usually meaning after you figure out something broke.