How to Use AI in Learning and Development (Without the Hype)

A year ago, it felt like every conference session, LinkedIn post, and vendor email was telling you the same thing: adopt AI now or get left behind. Some of you built a task force. Some of you quietly closed the tab and hoped the noise would pass. A few of you are still not sure this is a good idea at all, and that’s fair too.

The hype has mostly burned off. What’s left is a smaller, more useful question: how do you actually use AI in learning and development, day to day, without creating a mess you’ll have to clean up later.

What Actually Stuck

L&D’s adoption numbers jumped fast. The share of L&D professionals using AI went from about 5% in 2023 to 84% within about a year. That’s an adoption number though, not an impact number, and the two get treated as the same thing more often than they should be.

First drafts. Turning a blank page into a rough starting point, fast.
Summarizing. Pulling themes out of a pile of feedback nobody has time to read line by line.
Practice and roleplay. Low-stakes reps for conversations people usually only get to practice live.
Recommendations. Platforms quietly getting better at surfacing the right content for the right person.

One honest caveat: a widely cited study on AI-assisted work found people finished tasks slower by objective measures, while believing they’d finished about 20% faster. Feeling faster and being faster aren’t automatically the same thing.

▶

See practice and roleplay in action

Inside BizReady’s AI Roleplays: Where Practice Builds Skills

Watch the webinar

Where It Helps, Day to Day

Turning a messy brain dump into something sendable. You know what you want to say. Getting it onto the page in the right tone is the slow part.

Finding the theme in a stack of comments. Exit interviews, pulse surveys, open-text feedback nobody has time to read one at a time.

Prepping for a hard conversation before you have it. Talking points, anticipated questions, a rough outline before a tough one-on-one.

If you’re specifically looking at using AI to build out a full training program, we’ve walked through that in more depth already, prompts and all, in Designing an Employee Training Program Using ChatGPT.

Where to Still Slow Down

Recent research gives a useful frame here. Leaning on AI doesn’t appear to make people less capable, but it can erode confidence in their own reasoning. The deciding factor wasn’t whether someone used AI, but whether they passively accepted what it gave them.

✓Don’t paste anything you wouldn’t want in a public document.
✓Don’t accept the first draft as the answer. Revise it, argue with it, cut what’s wrong.
✓Check anything with a number or a fact in it, the same way you’d check a new hire’s draft.
✓Keep a human eye on anything learner-facing before it goes out.
✓Don’t let it make the call on a person. That stays yours.

How to Start Using AI in Learning and Development This Week

Here’s the smallest possible starting point: one task. Pick something low-stakes this week. Try it once. Read the output the way you’d read a first draft from someone new to the role. Decide if it earns a second try next week.

SHRM’s 2026 workplace research found individual contributors save close to four hours a week using AI, even after accounting for time spent correcting its output. In L&D specifically, teams already using it report two clear wins: faster production and a better experience for the learners on the other end.

Results will vary by role and task. Still, that’s a solid enough case for giving the one small task a real shot.

A Few Questions You’re Probably Asking

Do I need an AI policy before I can use any of this?

No. If your whole team is getting into it and you want to make it official, see the rollout guide above.

What shouldn’t I use AI for yet?

Anything evaluative about a specific person. AI can help you prepare for that conversation. It shouldn’t be having it for you.

Is this going to replace the parts of my job I actually like?

Reading a room, building trust, knowing when someone needs a different kind of support — not things a model does. AI is picking up the tasks around those things, not those things themselves.