Digital HR Leaders with David Green

The episode contains several genuinely non-obvious insights - particularly the compressed-hours reversal and its EMP impact, the control-tower governance model for AI, and the 'go first' principle for HR - but these are interspersed with generic transformation-speak, platitudes ('culture always eats everything for breakfast'), and vague futurism that dilutes the useful-per-minute rate.

“we found out that we had over 10,000 people working compressed hours... our EMPs went from plus 23 to minus 31”

“we have a control tower, um, that myself and the Jiku, so the executive member accountable for technology, um, manage month in and month out”

There are a handful of genuinely fresh structural choices - demanding Places as a non-negotiable scope item, reframing shared services as an 'experience hub,' using viral native-user observation to scale AI workflows - but the episode leans heavily on recycled HR frameworks (skills-based organisation, triangle-to-diamond, culture eats strategy) and the future-of-work section is largely received wisdom attributed to Google.

“I'm not coming if you don't give it to me... because I just sort of feel that strongly about it”

“we're trying to find those people and then what watch them and then figure out what they're doing and how and why and then scale that back into larger groups”

Sharon Doherty is a genuine large-scale practitioner with an unusually strong track record - HR and change director for Terminal 5 and Heathrow, a decade at Vodafone, Finastra CPO, now CPPO at an 80,000-person bank - she speaks from direct operational experience rather than theory, and her scope (Places budget second only to the CTO) is credibly senior.

“within 18 months 70% of the executive team were new. Within two years 60% of the top 300 new and by the end of this chapter... 40% of the organization new”

“I said I'm not coming if you don't give it to me”

The episode delivers a solid cluster of concrete figures - EMP swing from +23 to -31 and back, 10,000 people on compressed hours, 100,000 AI courses in Q1, 40,000 frontline users, 12 big bets, 50 first-year use cases, 70%/60%/40% leadership turnover milestones - though financial ROI, cost saves, and hard productivity metrics are almost entirely absent, and the super-agent section remains vague.

“in the first quarter 100,000 courses completed”

“50 use cases and then this year we started with 12 big bets and over 100 use cases”

The host shows genuine preparation - he cites prior conversations, references specific initiatives like the super agent, and connects the episode to prior research - but almost never challenges an assertion, frequently signals agreement before the guest finishes, and misses clear opportunities to press on metrics, trade-offs, or failed experiments.

“You've been busy”

“Very good”