I am a content junkie. Podcasts, books, articles, talks…if leadership or artificial intelligence is in the title, I am probably consuming it. And, most of it wraps up in the same place: When AI is doing the tasks, leaders may need to lean harder into human skills like empathy and emotional intelligence.
I agree, but I find myself frustrated with how this content is framed. It either predicts a future that has not arrived yet or speaks about that future as though it is already here.
Neither version matches my reality, which more closely resembles a Kübler-Ross rollercoaster of change, with every leader I work with in a different stage depending on the day. Some are energized, some are overwhelmed and some are quietly pretending they understand tools they have not had time to learn. Leaders are not struggling with how to lead after AI. They are struggling with how to lead during it—in the messy, constantly shifting middle.
This is where L&D leaders can help the most: in the middle. We should be actively creating programs that help leaders not only upskill but navigate this unprecedented transition for themselves and for the people they lead, while keeping the work and the business moving forward.
Lead self: Design for curiosity
What leaders are actually experiencing right now is not simplification, but accumulation. AI has not taken anything off of leaders’ plates. In fact, it has added a new layer of expectation on top of plates that were already full.
What complicates this further is that these new AI tools are moving faster than the systems around them. Expectations are rising faster than policies, support structures and shared norms, leaving leaders stuck projecting confidence in an environment that is still taking shape. That weight deserves to be addressed honestly.
And yet, the most common approach to AI adoption still leads with fear: Learn this, or someone who already knows it will take your job. That framing does not inspire curiosity. It triggers self-preservation, which does not foster an open, exploratory mindset that real adoption requires.
When people feel like they are being told to change or risk becoming irrelevant, they either comply at the surface or resist quietly, and neither outcome is what L&D is actually trying to build.
There is a better starting point, and it begins with cultivating curiosity, not fear. Instead of telling leaders they need to upskill, begin with a different question: What are the three things you dislike doing the most that are part of your job on a regular basis? Then, design workshops that teach them how to use AI to get those specific things done faster or automate them entirely. When you start there, the learning is immediately self-serving. It solves a problem the leader already wanted solved, resistance drops and willingness shows up all on its own.
This is how L&D can start offsetting the accumulation instead of adding to it. When AI adoption takes something off a leader’s plate, you change the emotional experience of the transition itself. Leaders stop bracing and start leaning in because they felt the relief firsthand.
Lead others: Meet people where they actually are
Here is what psychologist Abraham Maslow understood that most AI adoption strategies miss: People cannot grow when their foundational needs are unstable. Right now, in most organizations, the psychological basics are shaky. People are unsure if their role is safe, unsure if their skills still matter and not confident that asking for help will be read as curiosity and not incompetence.
You cannot layer innovation and experimentation on top of that and expect it to take root. And yet, that is exactly what most adoption playbooks ask leaders to do: Rally their teams around a future that their teams are not sure includes them.
The instinct for most leaders is to reassure: “It will be fine. We will figure it out together.” But vague reassurance in the middle of genuine uncertainty erodes trust faster than the uncertainty itself; people can feel the gap between what is being said and what is actually happening.
Additionally, no two people on the team are experiencing these changes the same way. Some are worried their role is next, others hear “AI will help you work faster,” and interpret it as more work on their plate and some are quietly mourning. Take the high-performer who used to love the analytical side of her job: the modeling, the data, the problem solving—and now she just prompts AI all day.
The output may be better or faster, but the craft is gone. Your high-performer is grieving a version of her professional self that no one told her she would have to let go of, and there is no single message in the current conversation that can offer security to your employees, all experiencing their own losses.
So, what does leading others through this transition actually look like?
It starts with declaring the middle out loud. When you are in front of a team that is clearly overwhelmed but not saying it, say it for them: “I know this is a lot. I know it is not clear yet. I am in it too.”
Then, ask real questions and stay for the answers; not in a survey, but in a conversation:
- “What has changed about your day-to-day work in the last six months that no one has officially acknowledged?”
- “What do you feel like you are supposed to say about AI, and what do you actually think?”
The leader who asks these questions and genuinely listens is modeling what human-centered leadership looks like right now: Present, honest and more interested in what is real than in projecting confidence. That is how you meet people where they actually are; where trust and forward movement can coexist. But even the most skilled, most present leader cannot sustain that work alone. At some point, the system has to meet them, too.
Lead the organization: The conditions for success only leadership can create
This is not just another change management challenge with an AI label on it. Every previous wave of workplace technology changed what people did. AI is creating an identity-level disruption and identity-level disruptions do not respond to the same playbooks as process-level ones. When leaders fail to navigate this well, people comply on the surface and disengage underneath, and the result is diminishing returns.
Recent research confirms what many learning leaders have long suspected. According to Microsoft’s 2026 Work Trend Index Annual Report, organizational conditions, including culture, manager support and talent practices, are more than twice as influential as individual capability when it comes to whether AI actually delivers value. Three barriers show up repeatedly, and none are fixed by simply adding more training.
The first is logistical. Governance takes time, security reviews are slow and the expectation that everyone “stay current” assumes a learning bandwidth most roles do not leave room for.
The second is cultural and largely unspoken. Many leaders are quietly embarrassed about using AI, unsure when it is appropriate to say “I used AI to do this,” and until they feel safe doing so without risking credibility, adoption will remain hidden and their teams will take the same cue. You cannot build a culture of experimentation on top of a culture of shame.
The third is an incentive contradiction. According to the same Microsoft report, 65 percent of AI users fear falling behind if they do not adapt quickly, yet only 13 percent say they are rewarded for experimenting with AI at work.
That is not an adoption problem. It is a design flaw. Organizations are telling people to change while continuing to measure and reward the old ways of working. This is where L&D needs executive partnership—not just executive permission—because programs alone do not move culture, but programs aligned with structural change can.
Each of these barriers has a corresponding move, and L&D is positioned to make every one of them. On the logistical side, L&D needs insight into provisioning discussions, or better yet, a seat at that table. If you do not know who has access to which tools, you will build training that will either confuse people or miss them entirely, and as those tools expand or contract, your audiences and content need to adjust in real time.
On the cultural side, the fix starts at the top. Leaders need to openly model and discuss where AI was used in their own work, not as a performance, but as a norm. Organizations need to share case studies of successful AI use so that experimentation becomes visible and celebrated. On the incentive side, innovation requires room for attempts that do not land, and that room has to be built intentionally with psychological safety in mind.
I have seen firsthand what happens when these conditions exist: People move faster, share more openly and treat AI as a tool worth exploring rather than a threat to outrun. That does not happen because of a training program. It happens because someone at the organizational level made deliberate choices about culture, incentives and infrastructure before asking people to change.
L&D is uniquely positioned to identify these barriers, advocate for the structural changes needed to address them and refuse to launch programs into environments that are not ready to support what those programs are asking people to do.
Lead from here
No one knows how long this middle will last, and that is part of what makes it so hard to lead through.
But the leaders and L&D teams who will prevail are not the ones waiting for clarity before they act. They are the ones inspiring curiosity when fear would be easier, meeting them with honesty when reassurance would be faster and pushing for the structural changes that make real adoption possible.
The honest middle is not comfortable, but it is where the most important leadership work is happening.
