Honestly, our first reaction to that was a little deflating.

If the problems haven’t changed, are we all just running in place? But the more we chewed on it, the more we think it says something different. These were never really technology problems.

Every new wave of tools, from the LMS to the year everyone was going to do VR, showed up promising to solve all our problems, and most of them did help with part of the work.

But the problems stuck around because they’re about people, time, priorities, and judgment, and no tool fixes those on its own.

AI is the most powerful tool our industry has ever had, which also makes it the most tempting band-aid to whip out of our first-aid kits.

But doesn’t it make more sense to figure out where the boo-boo is before we start slapping bandages on things?

Here are three conversations from DevLearn that stuck with us.

AI has made content cheap.

A course outline that used to take a week now takes an afternoon, and a first draft of a branching scenario can probably be done before lunch. Hey, that sounds pretty good, eh?

Welp, teams told us they’re producing more than they ever have, and learners are engaging about as much as they used to, if not less. That’s a big problem, friend-o.

We’re dating ourselves here, but when smartphones put a camera in every pocket, the world didn’t suddenly fill up with great photographers. It filled up with ten thousand photos of brunch and sneakers. And the dramatic selfies…my gosh, the selfies.

Taking a picture got easy, and the hard part, knowing what’s actually worth capturing, turned into the valuable part.

That’s kinda what’s happening to our industry right now.

When building a course costs almost nothing, the most valuable skill in the room is knowing when not to build one. When to put down the camera.

That means the bottleneck in L&D has moved from production to judgment, and a lot of teams haven’t rearranged themselves to match that new workflow.

The people who are good at saying “not yet” or “what if we just made a checklist?” used to slow projects down. Now they might be the most valuable people on the team.

Learners have also gotten very good at spotting AI-generated content.

We don’t have a study to point to, but we have a very strong hunch that some of what isn’t sticking isn’t sticking because it feels like nobody actually made it.

We’ve been noticing for a long time now that there’s an “othering” that happens when it comes to our learners. Like they’re mythical creatures living in the woods that we’ll never be able to truly understand. What COULD they be thinking?!

Hey guys, they’re us. They’re just us.

If you can’t sit through your own course without checking your phone, why should they? Did it make you feel anything at all, or did it feel like it was written by nobody, for nobody?

People can tell the difference between something created for them and something generated at them, and they tend to give the first one a lot more of their attention.

Ahh, this old chestnut.

This is the classic order-taker problem, and it’s alive and well. Someone senior has decided the issue is a training issue, and the solution is a 30-minute module, due by the end of the month.

It’s a little like walking into your doctor’s office and asking for the specific prescription you saw in a commercial. A good doctor should probably ask a few questions first, even if that medication has the catchiest jingle you’ve heard since “I Wish I Was an Oscar Mayer Wiener.”

AI makes this one trickier too, because now that course really can be built by the end of the month, and probably even sooner than that if they really wanna make it happen.

Which puts a lot of L&D teams in a Jurassic Park conundrum. Like Dr. Ian Malcolm said, “Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.”

A lot of what lands on L&D’s desk as a training request is really a process, tool, or incentive problem.

So why does everything turn into a course?

Partly because a course is the most visible way to show you did something about a problem. It’s something you can point at and say, “Hey, I did a thing! I tried!” A box to be checked.

Fixing a clunky handoff between two departments or a confusing piece of software is way slower and messier, and it doesn’t come with that nice little dashboard. Training becomes the default because it’s the easiest fix to point at, not because it’s the most likely to work.

And then of course, there’s the politics of it all. Pushing back against a senior request can feel like a career-limiting risk. So the order gets taken. And who could blame you in this economy?!

Gently pushing back doesn’t have to be “no.” Before anyone opens an authoring tool, try asking three questions:

• What should people be doing differently 90 days from now?

• What’s getting in the way of them doing it today?

• How will we know it worked?

In our experience, most leaders don’t actually want a course. They want the problem to go away. The course was just the only fix they knew how to ask for.

Which brings us to the question that somehow almost never gets asked. What are we actually trying to accomplish here?

Figuring that out is basically L&D’s superpower, and it’s the easiest way to push back without being a jerk about it. You’re not saying no. You’re helping the person who asked for the course get the win they were actually hoping for. And that just makes them look good too.

People like looking good. So, we've heard!

Of all the challenges we’ve helped clients tackle in 2026, this one comes up more than any other.

A huge share of L&D happens in teams of one or two, where the same person is the instructional designer, developer, project manager, LMS admin, facilitator, and whoever gets called when a projector won’t connect or something.

Sadly for us poor souls, that’s not new.

What is new are the huge expectations that come with it.

Somewhere along the way, “Now you have AI” turned into: