Me, Myself, and AI

The episode contains a handful of genuinely useful specifics - a three-tier AI deployment framework, measuring prompt volume by line of business as a cultural proxy, and the human-override 'sixth time' heuristic - but these are diluted by lengthy passages of corporate HR boilerplate about curiosity, collaboration, and 'humans in the lead.' Insight-per-minute is uneven.

“how many prompts an organization is writing by line of business. So what is the kind of culture of AI adoption by individual group around the organization”

“if you were using AI of sorts and it said yes, yes, yes, the key question has to be, well, what happens that sixth time?”

The framing is almost entirely conventional: AI is not a toy, humans must stay in the lead, soft skills still matter, don't fear replacement. The three-level taxonomy is tidy but not novel. There is no genuinely contrarian or first-principles argument made anywhere in the episode.

“AI is not good at judgment. That requires a human in the loop”

“it is not like a toy. Companies who treat that really seriously are going to start kind of top down”

Bernard Hampton is a legitimate senior practitioner running one of the largest corporate learning organizations in the world, with real authority and verifiable scale numbers; he has clearly done the thing. He is not a thought-leader or career podcast guest, though his role is L&D/HR rather than a P&L or technical operator, which limits the density of hard operational insight.

“45% of open roles last year almost were filled internally. Which further leans into why skilling and upskilling are so important”

“we have one that clients use more than 169 million times a quarter and growing quarter after quarter”

The episode supplies a solid set of organizational metrics - headcount, internal fill rates, financial center count, client AI usage volume - which anchor the conversation in real scale. However, there are no outcome metrics for the learning programs themselves: no improvement in time-to-proficiency, no productivity uplift numbers, no before/after data on simulation effectiveness.

“20,000 people we hired last year. That includes 2,000 of our student campus hires. 45% of open roles last year almost were filled internally”

“clients use more than 169 million times a quarter and growing quarter after quarter”

The host occasionally surfaces a real tension - most notably pushing on the risk that KPIs will bias organizations toward easy Level 1 efficiency wins and away from transformational Level 3 work - but mostly lobs open-ended prompts and lets vague answers pass unchallenged. The rapid-fire section and the Good Place digression eat time without generating substance.

“I worry that the prevalence of those sorts of measures are going to lead us towards really focusing on what you call that level one, which is much easier to measure”

“Are there things that you've looked at on paper that hey, this might be a good place to use AI, but you've decided that the risk didn't pay out”