The Edge of Work

The episode contains a handful of concrete implementation details—Lean Six Sigma applied to AI rollout, a structured ambassador program, a functioning HR chatbot—but the surrounding conversation is heavily padded with change management platitudes and repetitive motivational framing that dilutes the useful-per-minute ratio significantly.

“we actually certified people in lean Six Sigma methodology so that we would have it viewed as a more of a project management change management initiative”

“almost 85 to 90% of the questions coming through it as like tier one and two, tier two level questions. That were coming in being answered with very high levels of certainty and accuracy”

The 'co-parenting AI between HR and IT' framing and the Lean Six Sigma application to AI transformation are modestly fresh angles, but the core arguments—AI shouldn't replace jobs, humans in the loop, bottom-up beats top-down—are widely circulated takes that appear in virtually every enterprise AI adoption conversation.

“AI is, we're co parenting it between technology and the HR group, we're raising it together as our baby”

“we decided we were going to use AI as co workers on the team to serve a specific purpose and take some of the administrative, repeatable tasks and burden off”

Robin Barbacane is a genuine practitioner who has actually designed and shipped specific AI programs at a named technology company, which is meaningfully better than a generic thought-leader; however she is a mid-to-senior HR function lead rather than a P&L owner or C-suite executive, and some of her framing skews motivational rather than operational.

“for 30 years I've been in HR. For the last 15 years, I've been working on large scale transformation”

“We knew within our organization, we set out from the beginning some certification and basically code of conduct or ethics for our organization”

There are some genuine program-level specifics—chatbot handling 85-90% of tier-1/2 queries, 495 ambassador applicants whittled to 17, a six-week sprint cadence, $20-40/month licence cost framing—but there are zero hard ROI figures, no dollar savings cited, no productivity improvement data, and no named vendor or tool implementations beyond a casual Claude mention.

“almost 85 to 90% of the questions coming through it as like tier one and two, tier two level questions”

“We had 495 people raise their hand... Only 17 people were selected”

The host structures the interview logically and asks a few genuinely useful process questions, but consistently validates and editorialises rather than probing—no numbers or claims are challenged, vague outcomes go unquestioned, and a mid-episode self-promotional break disrupts momentum entirely.

“I think that's a really fascinating approach in terms of not only does this look a little different in terms of at least stacking this up against other AI transformations that I've witnessed or seen”

“has it been mostly like other transformations that you've done in the past? Is it different?”