I Owned L&D and Here's What It Doesn't Do...yet

For most of human history, knowing something and teaching something were the same act. Skills lived inside people, and the only way to move them was to put the right bodies in the same place to converse. Two people in a cave.

The printing press ramped up scale. And the internet collapsed distance entirely. Each shift solved the same underlying problem: how do we consolidate knowledge and get it to more people faster?

From there, incremental improvements. MOOCs. LMS platforms with completion tracking. Wikipedia. YouTube, weirdly enough. The internet had all the information, and different platforms made it marginally easier to consume on demand.

But every incremental improvement assumed the same thing. Learning is a distribution problem. Find the right content, get it to the right person, verify they consumed it. That model made sense when content took time to organize and distribute.

It’s the wrong model now.

Which just created a different problem. There’s too much information available. There’s real work in knowing what you actually need, and when.

But set that aside for now because we can solve that while we solve this one…

Behavior change.

At its core, this main issue remains largely unresolved.

Every shift in how we move knowledge changed the distribution. Writing, print, the internet. None of it changed how humans actually adopt new behaviors. The forgetting curve is the same. The knowing-doing gap is the same. A completion certificate has never made anyone meaningfully better at their job.

AI is the first tool that can show up at the moment the behavior is actually happening. Not before it in a training, or afterward as feedback. Right then.

But most of the L&D industry is still focused on producing content faster and with better production value. We have the first real shot at closing the gap, and a lot of people are using it to build slicker slide decks.

If that’s “bad”, what does good look like?

Picture a standard training on company values. A facilitator, some employees, and a minimalist slide deck Steve Jobs would be proud of. All in a room. That’s it.

Afterward, back at work, you’re writing an email to a frustrated client. As you go to hit send, something surfaces on screen…feedback for behavior change in the moment.”This lands as defensive. Here’s one way to reframe it that’s more consistent with our values while having customer interactions.”

You didn’t ask for it. The training from last Tuesday wasn’t on your mind at all. But the behavior it was trying to change showed up anyway, right before the damage was done.

Not better content. Better timing.

The strange part is this isn’t a new idea. Follow the same idea back through history and we land where we started this article. With two people in a cave.

You learned to start a fire by standing next to someone who already knew how, and he stopped you in real time when you went for wet wood. No training. No post-assessment. Just feedback at the exact moment the behavior was happening. That model produced some of the most durable skill transfer in human history.

We didn’t abandon it because it stopped working. We abandoned it because it didn’t scale. You can’t put a master practitioner next to every employee. So we built the next best thing. We wrote the knowledge down, packaged it, distributed it. And we’ve been optimizing that compromise ever since.

L&D has been citing the 70-20-10 model in strategy decks since the 90s. Seventy percent of learning happens on the job. We’ve known for a long time that real-time feedback beats pro- or retroactive training. We just had no practical way to deliver it at scale, so we kept building courses instead.

Now the tool exists.

So the question is, what are you building with?