For the better part of two decades, the conversation about technology and HR has run on a single anxious loop: the machines are coming, and they’re coming for us. Every new wave arrived wrapped in the language of disruption, from the applicant tracking systems that promised to fix hiring to the analytics dashboards that promised to predict everything to the generative models that can now draft a policy in the time it takes to refill your coffee. The framing was always the same, as though HR were a taxi medallion and AI the rideshare app circling the block, headlights on and meter running.
But disruption is the wrong word, and on reflection it was always the wrong word. AI isn’t the disruption for HR. It’s the diagnostic.
It’s the moment someone turns the lights on at closing time, and now you can finally see what the bar really looks like – the sticky floor, the duct tape on the third stool, and the patron who has been nursing the same flat drink (and the same grievance) since four o’clock.
Nothing has changed about the room; it’s just that the lights came up and what had been comfortable in the dark is suddenly – unflatteringly – visible.
Hiding in the Dark
For a very long time, HR’s professional identity rested on “volume”. We were the keepers of the information, the processors of the forms, the people who knew which policy was which, how folks enrolled in their benefits, and which employee was three documented incidents into a problem.
This gatekeeping, while decidedly unglamorous, wasn’t trivial; it was real work, and it made us necessary. But necessary is not the same as strategic, and the sheer mass of administrative throughput gave the profession a place to hide whenever someone asked the harder questions about our value and contributions.
AI can now remove that volume. It can do the transaction, the routing, the first-draft summarizing, and the predicting-attrition-with-colorful-graphs that we used to do by hand and call insight. And once the transactional layer is gone – once the machine has absorbed everything that was process pretending to be judgment – what remains is precisely the work we always claimed was the point…
… the strategy, discernment and the human judgment about people, power, and consequences. The stuff we have long said no machine can replicate on our behalf.
Transactions to the machine, and judgment to the human. Which sounds mighty fantastic as a slogan, but also sounds a bit brutal as a mirror. Because it leaves nowhere to stand for anyone who built a career on the transaction piece and quietly hoped the judgment bits would never be audited.
You Can’t Stay Here
The thing about closing time is that you don’t get to vote on whether the lights come on. They get flipped on regardless of your desires and the only real decision left is what you do once you can see. You can grumble about the glare and insist the place looked better in the dark – which is true but also beside the point. As the song goes, “you don’t have to go home, but you can’t stay here…”
So with AI being the diagnostic for HR, it’s time, when those lights go on, to take an honest look at what the diagnostic is offering and decide what you want the HR function to look like in the morning.
The HR leaders I’d put money on are the ones treating this less as a threat to be managed and more as a long-overdue reckoning to be used; the ones moving toward something closer to a human capital consultant than an administrator. They’re the HR professionals advising on the issues that matter – because they refuse to be reduced to a workflow.
But they’re not waiting to see whether the lights are flattering. They already know they won’t be; the “function” the lights are illuminating has, perhaps, never been as strategic as we always insisted it was.
And that is why disruption was always the more comfortable story to tell. Disruption is something that happens to you – an external force with no opinion about your competence, and a convenient one to blame for the wreckage – while a diagnostic implicates you. It doesn’t ask whether the technology is coming; it asks what that technology finds when it arrives. And, more pointedly, it asks whether the function it illuminates was adding value all along, or was simply well-lit enough, until now, to never have to find out.
*****
