Drawing the Line: A Leader’s Guide to Human-Centered AI Adoption

A few months ago, a CHRO asked me to step into a situation that, on the surface, looked like progress. The organization had invested heavily in AI tools, automated large portions of its HR processes, and positioned itself as an early adopter. But beneath that momentum, something wasn’t working.

In hindsight, the CHRO was clear: they had moved too fast. Like many leaders today, they felt pressure from various sources to act quickly and not fall behind. So they leaned in. In this case, that meant automating nearly the entire onboarding experience.

What they didn’t fully anticipate was what that decision would signal…and how it would feel.

Even with a ‘human in the loop’, new hires were consistently directed to an advanced chatbot for questions and guidance. And technically, it worked, as it was operationally efficient. But experientially, it fell short. People didn’t want efficiency in that moment, they wanted connection and the ability to be seen.

The data reflected it almost immediately. Onboarding experience scores dropped. Early-tenure attrition rose, with more employees leaving within their first 90 days. And the feedback was strikingly consistent: being routed to AI during such a critical phase made new hires feel like an afterthought.

The consequences went beyond experience metrics. Financially, the organization absorbed the cost of replacing employees who left early, which was no small figure. According to SHRM, the average cost-per-hire exceeds $4,700, with total replacement costs often significantly higher when factoring in lost productivity and ramp-up time. Culturally, the impact was just as real, as hiring managers grew frustrated as roles reopened, and trust in HR began to noticeably erode.

This is where many organizations are finding themselves. According to recent research, more than 80% of HR leaders are actively exploring or implementing AI in their functions. At the same time, Deloitte’s research shows that employees continue to place increasing value on human-centered experiences at work. Those two realities are not in conflict, but they do require thoughtful navigation.

What’s often missing in these moments is not capability, but judgment.

Too many AI adoption decisions are framed as efficiency plays when they are, in fact, decisions about risk, culture, and accountability. When you delegate a process to AI, you are not just changing how the work gets done. You’re changing what that moment means to the person experiencing it and who owns the outcome if something goes wrong.

That’s the line leaders need to see clearly: the human/AI delegation line.

In the case of this organization, the turning point came when they stepped back and asked a different set of questions. Not “What can we automate?” but “What should we automate?”.

They reintroduced real touchpoints into onboarding such as manager-led conversations, live check-ins, and intentional moments of connection. AI didn’t disappear, but it was instead repositioned. It handled administrative tasks, surfaced insights, and supported consistency, no longer replacing the moments that mattered most.

The results were decisive. Onboarding scores rebounded, early-tenure retention stabilized, and perhaps most importantly, trust began to rebuild.

There’s a broader lesson here for leaders navigating similar decisions. Before delegating any meaningful process to AI, it’s worth pausing on a few fundamental questions: Who is truly accountable if this goes wrong? What are the legal and regulatory implications? What message does this send to employees about what the organization values?

And another equally important consideration is to question whether the right people were in the room when the decision was made. Too often, these choices are driven by vendors or technology teams, with HR, legal, and employee experience perspectives brought in too late (or not at all!).

And then there’s the simplest, most human question: does the person on the receiving end of this process need speed, or do they need to be seen? The answer is not always the same, but it is always consequential.

None of this is an argument against AI. In fact, it’s quite the opposite. When applied thoughtfully, AI can elevate HR’s impact, freeing up time and improving decision-making to enable greater consistency at scale.

But the organizations that get the most value from AI won’t be the ones that move the fastest. They’ll be the ones that are most deliberate about where they draw the line.

If you would like to discuss how we can help your organization implement AI so that it is used towards the long-term sustainability of your business, please get in touch with me at carlos.butler-vale@orgshakers.com