"We are pushing all of this AI experimentation, but yet, we are not reducing our performance metrics." — Adrienne Guerrero, Positive Delta

That quote named something most organizations haven't said out loud yet.

People leaders from CNA Insurance, Combined Insurance, Spencer Stuart, UL Solutions, The Regis Company, and across the ELE community gathered to work through a problem hiding in plain sight: employees aren't keeping skills current — not because they don't care, but because no one's adjusted what gets measured. Perform now, learn later is the rational choice. Until that changes, upskilling is just noise.

Who was in the room: Nicole DeFalco facilitated. Vilson Simon, Adrienne Guerrero, and Mike Hruska led breakouts. Matt Sernaker produced. Baryons AI participated live — synthesizing room signals in real time.

The reframe that stopped the room

Three breakout perspectives — HR business partner, frontline manager, employee — surfaced the same signal: no one's been given a clear reason why upskilling matters or what's in it for them. Managers aren't being villains. But they're signaling, unintentionally, that delivery wins.

BARYONS AI Mentor fed it back:

"If we reframe upskilling as a funded investment in job security, rather than a threat to productivity, we can convert employee fear into innovation."

That wasn't a framework from a slide. It was the room's own thinking — sharpened. The smiles weren't about the AI output. They were recognition.

What the room produced

A 30/60/90 scaffold: communicate the why in 30 days, adjust metrics to create safe learning zones by day 60, formalize a reward structure for both learning and output by day 90. That thread moves into the AI Practical Lab on June 11 for deeper implementation work.

Mike Hruska's breakout tested an approach worth stealing: the AI human sandwich — human framing first, AI synthesis in the middle, human challenge and iteration at the end. Marty Murrillo put it simply:

"You have to challenge it, think, and iterate. You're the director."

That skill — challenging AI output before you act on it — isn't in most upskilling plans yet. It should be.

The open question nobody answered

Azizeh Constantinescu, UL Solutions, asked the one that landed hardest:

"When AI can do the task faster, are we cutting off human growth and development at the knees?"

Classic change models assume a destination. AI doesn't offer one. Any reskilling strategy built toward a defined endpoint will be obsolete before it's deployed. The leaders getting traction are designing for iteration — not installation.

---

June 11, the same co-presenters take this into coaching workflows -- live. If you missed June 5, this is where the work continues.

Join the June 11 AI Practical Lab: https://www.ele.llc/calendar/ai-practical-labs-23