Dear L&D,

A few weeks ago, I was speaking with an internal trainer at a large bank. She’d just finished facilitating a change management workshop when a participant asked:

“Why would I spend half a day in a workshop when Copilot can explain this in five minutes?”

The trainer didn’t get defensive, but she didn’t have a ready answer either, so she reached out to me. The more I sat with her question, the more I realized it wasn’t really about artificial intelligence (AI). It was about what organizations expect when they invest in learning.

The question felt familiar because it echoed another moment in learning and development (L&D) history. I was early in my career when the eLearning wave hit. Like many people, I thought, “This is incredible. Now anyone can learn anything, anytime!” Around 1999 and 2000, I genuinely believed people would wake up on Sunday morning excited to learn about change management in their pajamas.

It all seems silly now, but at the time, those predictions felt inevitable. Looking back, it’s easy to smile at how overconfident many of us were.

The Last Time We Thought Everything Would Change

E-Learning promised, or depending on your perspective, threatened, to replace facilitated classroom learning. It was the future. And in many ways, it was transformative. It made content accessible at scale and fundamentally changed compliance training and onboarding.

But it didn’t replace facilitated learning altogether, and it didn’t live up to the hype of those early days.

Today, AI is the future. In many ways, it has delivered on eLearning’s promises. It has further commoditized content. It has also commoditized personalized context: With a few minutes of prompting, you have a coach who’s familiar with your situation, your colleagues and your calendar. That would have been the stuff of science fiction in the early days of eLearning.

AI is even starting to commoditize learning methods such as scenario generation. A scenario-based practice activity could take a specialized learning company months to build. Now AI can draft one from a prompt.

So What Becomes Scarce When Anyone Can Learn Anything?

Imagine you are facilitating a team through a leadership tool such as a Force Field Analysis. Before the workshop starts, someone says, “Let’s save some time,” and asks ChatGPT to generate one. Five minutes later, it is done. It looks polished, better organized and more confidently written than anything the group would have produced.

Yet you still want to throw it away because the analysis was never the point. The work was not producing a Force Field Analysis. It was watching someone say, “I don’t think that’s what’s holding us back,” and seeing three heads nod around the room. It was discovering that two teams had completely different assumptions about the elephant in the room. It was realizing the resistance was not where anyone thought it was.

That is what so many debates about the death of facilitated learning miss: The conversation is the work. The flip chart covered with sticky notes is evidence that it happened.

None of this means we should keep AI outside the room. AI can help people prepare, challenge assumptions and turn discussion into action. Used well, it can make those conversations more focused and productive.

AI may be reminding us of something we should have known all along: Good facilitation creates the conditions for people to question, practice, disagree, make sense and move forward together.

As Knowledge Gets Cheaper Alignment Becomes More Valuable

AI is extraordinary at helping one person reach “I know.” But AI alone cannot get a group to “we agree,” and that agreement is vital for teams to move together, quickly. That’s why organizations invest in learning in the first place: not simply to transfer knowledge, but to build shared understanding, language, expectations, judgment and norms.

A talent leader I respect once told me, “Culture is built through shared experience.” That’s why the answer to the participant’s workshop-versus-Copilot question isn’t, “AI can’t help here.” Clearly, it can. What it can’t do on its own is create the conversation that changes how a team thinks, decides and acts together. The ability to move people together matters more than it used to. When change is constant, an organization’s competitive advantage may depend less on individuals adapting on their own and more on teams adapting together, or what we might call communal agility.

Three Tradeoffs Worth Watching

If communal agility is the real advantage, we should ask whether our enthusiasm for AI is pushing us to optimize for the wrong outcomes. Three tradeoffs come to mind:

1. Personalization over shared experience.

AI makes personalization inexpensive and widely available. But not every learning challenge is individual. Sometimes the greatest value comes from having everyone wrestle with the same problem, hear the same conversation and leave with the same language. Personalization can make people individually smarter but collectively less aligned. Shared understanding helps people pull in the same direction.

2. Efficiency over shared sensemaking.

AI can produce answers almost instantly. But faster answers do not automatically make organizations more capable. Capability grows when people wrestle with ambiguity together, challenge assumptions and build a shared understanding of what to do next. Shared judgment helps people make consistent decisions.

3. Individual productivity over collective capability.

AI can help individuals perform better. Organizations, however, succeed when teams perform well together. As AI raises individual capability, collective capability may become even more valuable. Shared capability helps people work effectively together.

A Question to Ask Before You Book the Room

All of which brings me back to the participant question that started this letter: “Why would I spend half a day in a workshop when Copilot can explain this in five minutes?”

When that question comes up, here’s my new go-to response: If one person learned this perfectly and nobody else did, would the problem be solved?

If the answer is “yes,” consider AI-supported learning. It can work well for technical instruction, private practice, personalized coaching, preparation for difficult conversations and compliance training.

If the answer is “no,” ask why. Usually, one of three things is true:

1. The problem requires shared understanding.

Do we all understand what we are trying to do? The harder the problem, the more important it is for people to develop a shared understanding of what matters and why.

2.The problem requires shared judgment.

Do we make similar calls when the situation is not clear? When the answer is often “it depends,” people need a shared approach to making sound decisions.

3. The problem requires shared capability.

Can we do what needs to be done together? Solving problems, making decisions and collaborating require shared skills and ways of working.

So when that question comes up again, and it will, we can acknowledge what AI does extraordinarily well and explain why learning together still matters. Organizational learning isn’t only about helping one person know more. It’s about helping a team develop shared understanding, judgment and capability.  That’s how teams can quickly adapt, make consistent decisions and execute well together.