The Most Human Thing AI Can Do Is Get Out of the Way

Every few months, a new survey lands confirming what most leaders already feel in the room: employees are anxious about AI. They worry about relevance, about being replaced, about what their jobs will look like in three years. And paradoxically, some of that anxiety is pulling people away from the human connections that would actually help them navigate it. They’re distracted and only semi-present, waiting to see what happens next.

What if we’ve been looking at the problem the wrong way?

The conversation about AI in the workplace tends to focus on efficiency. The metrics we’re looking at sound like “fewer hours spent on X,”” faster turnaround on Y,” or “cost savings on Z.” That framing isn’t wrong, but it’s incomplete. Sure, AI saves us time and it can impact the quality of our most human interactions.

The most powerful way AI can develop talent is to subtly guide and support human relationships.

The 45 Minutes Problem

Here’s a scene that plays out in organizations every day: a mentor finally can carve out 45 minutes to meet with their mentee, and they arrive at the meeting room with a coffee in hand. The mentor asks, “What did we agree on last time?” They spend the first 10 minutes hunting for notes and getting distracted from the original intention of the meeting. By the time the conversation actually starts, a third of the meeting is gone.

Now multiply that by every developmental conversation in your organization. Every mentoring relationship, every coaching check-in, every career conversation that someone booked with good intentions and showed up to underprepared. And the problem is rarely the people; the problem is the administrative work required to get the most out of the conversation.

The numbers reflect this cascading problem with development conversations. A May 2024 Gartner survey of more than 3,375 employees found that nearly 6 in 10 said they aren’t getting the on-the-job coaching they need to support their core job skills. In other Gartner research, three-quarters of HR leaders reported that their managers are overwhelmed by the expansion of their responsibilities. Organizations value development, but it keeps getting crowded out.

When someone spends mental energy reconstructing context, tracking what was committed to, figuring out what to raise today, they cannot be fully present. And presence, it turns out, is most of the point.

Why Presence Isn’t a Soft Metric

There’s a category of human capability that doesn’t develop through content consumption. You can watch a hundred videos on ethical reasoning, on navigating ambiguity, on knowing when to trust your gut and when to slow down. The actual capability gets built through being in the room with someone — really in the room — through conversation with someone who challenges your framing, names what they observe in you, shares how they’ve handled something similar. Through the particular back-and-forth of a relationship that has context and continuity. That only happens when both people are fully present.

Judgment, sense-making, and identity development are the capabilities that will matter most as more routine work gets automated. They’re also the ones we’re most likely to underinvest in, not because we don’t care, but because we schedule the conversations and then show up with half our minds somewhere else.

SHRM’s research on workplace mental health found that opportunities for growth represent the single biggest factor in employees’ overall well-being, surpassing even job security. 76% of employees consider mentoring vital to their career development, and organizations with strong coaching cultures see real gains in productivity and engagement. People want these conversations and show up for them. And then the meeting starts with 10 minutes of “wait, where were we?” and the time slips away before they can make use of it.

Presence is what makes developmental conversation actually develop something. When it’s gone, no amount of good intentions fills the gap.

AI’s Best Move Is Supporting Human Connection

AI is good at the things that drain presence. It can help prepare context before a conversation, capture what was said after one, track commitments, surface the thread from last time, or flag what’s been left unresolved. None of these tasks require human judgment, and all of them currently consume it.

When AI absorbs that cognitive load, something subtle happens people look up. They make eye contact and the conversation starts from a shared, accurate understanding of where things stand rather than a negotiated reconstruction of the past. The 45 minutes becomes 45 minutes.

This is a different argument than “AI will make us more efficient.” It’s closer to “AI will help us focus and be present,”handling more of the information processing, which makes the human relationship more valuable, not less. The mentors and managers who show up prepared and present aren’t being replaced by AI, rather they’re being freed up to do the thing only they can do.

A good mentor doesn’t just transfer knowledge. They notice what their mentee isn’t saying, and they ask the question that reframes everything — those are the magic, career-changing moments in mentorship. None of that happens in a meeting that’s still warming up at the 10-minute mark.

A 2024 HBR piece on executive presence found that what employees now expect from leaders has fundamentally shifted, with a “listen to learn” orientation and authenticity rising sharply alongside traditional markers of gravitas and communication skill. Presence has become more important even as it’s harder to sustain. AI doesn’t have to threaten that, and when used well it protects it by getting out of the way.

This Is a Culture Question, Not a Tooling Question

The technology is not the hard part.

AI tools that support developmental conversations exist, but the harder question is whether your organization has the culture to adopt them in a way that actually serves people. Research published via MIT Technology Review found that fewer than 4 in 10 leaders rate their organization’s current level of psychological safety as “very high,” and that psychological barriers are proving to be greater obstacles to enterprise AI adoption than technological challenges. When people are anxious and norms are unclear, tools are avoided or misused and they become another source of distraction rather than a reduction of it.

If your managers are nervous that using AI to prepare for a conversation makes them look lazy, they won’t use it. Meanwhile, if employees feel like AI-captured notes are surveillance rather than support, they’ll resist too. Getting the technology in place is only 10% of the work. The other 90% is building the conditions where people trust it, understand its purpose, and have agreed on what it’s for.

Leaders set that tone. Gartner has found that organizations need to set clear human-AI collaboration norms specifically to reduce employee anxiety about job preparedness. Further, leaders who communicate openly about how roles will evolve with AI are better positioned to retain the talent they need. Every signal about how AI is expected to fit into the work of developing people shapes whether teams experience it as a threat or a resource.

Simply helping employees adopt AI will only lead to an increase in productivity (i.e.,” fewer hours spent on X”), but experiences where AI is intentionally integrated to guide and support relationships can help employees be more present and develop them in the way that only other humans can.

Two Paths Forward

Every company is making a decision right now about what AI is for. Most aren’t making it consciously. The default accumulates … tools get adopted here, use cases expand there, norms never quite get articulated, resulting in a culture that has formed but that nobody intentionally designed.

These are the two directions that culture can take.

One path treats human conversation as a cost. In this path, AI simulates the relationship: it generates development plans, delivers feedback, substitutes for the manager who doesn’t have time. The efficiency gains are real in the short term but the cost (slower but also real), is a workforce whose most important capabilities atrophy because the conditions for building them have been slowly dismantled. Gartner’s own research flags this risk directly, predicting that as AI flattens organizational structures and reduces middle management, mentoring and learning pathways may break, and more junior workers could suffer from a lack of development opportunities.

The other path treats human conversation as the asset and uses AI to support it. AI handles the prep, the capture, and the follow-through. The conversation itself stays human, but the manager arrives knowing the context and the mentee leaves knowing what was decided. The relationship compounds over time instead of restarting every 45 minutes.

The second path isn’t softer or more sentimental, it’s more strategic. Gartner predicts that by 2027, half of enterprises without a people-centric AI strategy will lose their top AI talent. The capabilities that will differentiate organizations in the next decade are built in the second kind of culture, not the first.

Getting out of the way requires restraint: a deliberate choice to use AI for the logistics and leave the relationship alone. But that restraint is exactly what makes the human moment possible.

The companies that figure this out won’t necessarily have the most AI. They’ll have people who are most present with each other. Right now, that’s a competitive advantage hiding in plain sight.