What if the ideas that L&D has been nodding at for thirty years are finally about to become unavoidable? Bob Mosher has spent his career arguing that training and performance are not the same thing — and that building courses, however well-designed, only meets two of the five moments when people actually need to learn. The other three happen in the workflow, at the point of need. Most of the profession has agreed with him in theory. Rather fewer have changed what they do.

In this conversation, John and Bob catch up on how generative AI is changing that picture. Bob describes what AI is actually doing inside live workflow learning projects right now — including a headline figure on development time that stopped John in his tracks — and explains why the digital coach, not the course, is becoming the central deliverable. They also dig into the risks: what happens when AI accelerates the wrong thing, why L&D needs a seat at the IT table before it's too late, and what it means to be a performance architect rather than a training order-taker in an AI-assisted world.

The boulder, Bob says, has been going uphill for a long time. He thinks it's about to roll back. This episode is about what happens when it does.

TIMESTAMPS

00:00 - Start

01:55 - Intro

04:27 - The five moments of need

08:14 - How has AI changed the conversation about workflow learning?

10:44 - How does AI save time?

24:07 - Risks to workflow learning from AI

31:42 - Has AI caused a rethink in practice?

40:19 - Workflow learning impact

53:38 - End

CONNECT WITH LEARNING HACK

LinkedIn: https://linkedin.com/in/johnhelmer

X: https://x.com/johnhelmer

Threads: @jphelmer

Bluesky: @johnhelmer.bsky.social

Website: https://learninghackpodcast.com