Learning Uncut

The intake methodology questions are genuinely useful and non-obvious, providing a concrete framework practitioners can apply immediately. However, the episode spends significant time on career autobiography and closes with vague AI speculation and motivational platitudes, diluting the density of actionable insight.

“My first question was always, tell me about your learners. My second question was always, will they be happy to take this training? And about 70% of the time the answer was no. And then I would let that kind of just go for a second and then I would lean in across the table and I would say, then why are we doing it?”

“let's imagine this training is so successful that the board of directors invites you to speak to them about how this is the most spectacular training the company has ever seen. But the catch is you only have five minutes and one slide. What's on your slide?”

The intake questioning framework has genuine craft and the 'wisdom management' framing for AI-era L&D is a novel angle, but the Bloom's Taxonomy revision is mentioned without enough detail to evaluate, and the broader AI discourse mirrors widely circulating L&D conversation. The business-professional-who-uses-L&D-tools reframe is common.

“I posted an article, um, with uh, a suggestion of a completely new set of verts for Bloom's Taxonomy. So these would include verts like prototype and govern”

“knowledge is not just what's documented. It also um, happens in the artifacts of our daily work. It happens in the tacit knowledge that's stuck in our experts heads”

Greg is a genuine practitioner with real scale credentials - 50,000+ users, 24-language global programmes, financial advisor onboarding for thousands of candidates, and an early enterprise generative AI use case - making him credible and relevant. He is not a prominent named executive or widely published researcher, keeping the ceiling moderate.

“I've led learning campaigns for over 50,000 users now”

“I also got the opportunity to uh, work on the firm's first ever generative AI use case. Um in 2023 our use case was selected um, out of a pool of candidates”

There are real numbers - 70% negative learner sentiment, 95% estimate accuracy, 50,000 users, 24 languages - and the intake questions themselves serve as concrete artefacts. However, the employing companies are unnamed, the Anthropic report is referenced only in passing without data cited, and the AI future section is largely abstract.

“about 70% of the time the answer was no”

“we got uh, to about a 95% accuracy rate for our estimates”

The host uses short, effective prompts ('Could they answer that question?', 'Tell me more') that do draw out Greg's framework, but there is no meaningful pushback, no challenging of claims, and too much airtime given to the host validating her own aligned views rather than pressing the guest further.

“Could they answer that question?”

“Tell me more.”