Contributor Thoughts: What CEOs Still Get Wrong About AI
ABOUT THE AUTHOR: Michael Anton is CEO of Huper, an AI startup based in Atlanta whose platform gives every leader a digital chief of staff.
The productivity gains we’re seeing from AI are very real, but the problem is those gains aren’t showing up in the places that matter. A recent survey found that 70% of executives reported productivity improvements from AI, but only 19% saw any lift in revenue. That gap is the real story.
We’ve spent the last few years asking whether AI would replace workers, when we should be asking what happens when each employee on a team triples their output overnight while leadership has the same bandwidth. The short answer is that leadership becomes the bottleneck. AI accelerates execution, but most leadership structures haven’t evolved to absorb that acceleration. That mismatch is where the real strain shows up.
What gets lost are the things people think are optional: the quick check‑ins, the mentoring moments, the casual walks that give everyone a little breathing room. But that’s the real people work. Skip it, and you stop developing your team, you stop building trust, and you end up with a group sprinting in the wrong direction because no one is slowing down to define what the work should actually be. Those “optional” moments are the glue that prevents misalignment, churn, and poor decision‑making. Without them, organizations drift fast.
The post‑AI world shifts the job from simple execution to real judgment. The tools handle the tasks, and people handle the thinking. That mindset frees leaders to focus on what actually matters and gives employees the chance to build the skills that move them forward. The real crisis is that critical thinking has been squeezed out of early careers, leaving new workers with no visibility into how decisions get made.
Half of C‑suite leaders already report de‑skilling in their organizations, and more than 60% expect it to become a material threat within five years. If companies want a real talent pipeline, they have to redesign how mentorship works and bring it back to the center of the job. Judgment work looks like prioritizing, interpreting nuance, making tradeoffs, and understanding context. These are the parts of the job AI can’t do for you.
Here’s what modern leadership looks like in three big shifts. First, stop asking to be copied on everything and get clear about the few moments that actually need your attention. Second, delegate real judgment, not just tasks, so people have the authority to decide how to handle problems instead of waiting for instructions. And third, build trust by checking in less, not more, so the team has room to think and move without constant oversight.
When teams work this way, everything moves cleaner and faster. Managers aren’t drowning in updates, leaders aren’t overwhelmed, and employees actually get the autonomy and mentorship they’ve been missing. Status meetings naturally shrink to fit the new flow. And once an organization has a real‑time view of what’s happening in the work, all those constant check‑ins and update layers become noise you can drop. What’s left is the communication that actually matters: strategy, escalations, tough decisions, and conversations with customers. AI is the mechanism that gives leaders that real‑time visibility, but only if they redesign how information flows.
Leaders shouldn’t be drinking from a firehose. Nor should they serve as a bottleneck. The CEOs who treat AI purely as a cost‑cutting tool are going to lose to the ones who treat it as human‑capital infrastructure and see the opportunity for better operators, better retention, and a functional talent pipeline.
The economics might be the same either way, but the outcomes are not. One future uses AI to amplify people, the other uses it to replace them. Only one of those futures builds a stronger company.
