"For leaders, what we need to be is responsibly curious. Ask questions, be interested, not interesting." — Michael Grubich, LAK Group

The real work in the room

Senior talent leaders, executives, and academic advisors came together to move past the hype and get honest about what AI adoption actually requires — not just tool access, but structural redesign. The work was getting stuck where organizations were overlaying AI onto unchanged roles, skipping the governance architecture that builds real trust, and treating adoption as a communications problem rather than an organizational design problem. It matters now because the window for deliberate design is narrowing: the organizations pulling ahead aren't moving faster — they're moving more intentionally.

Leaders in the conversation

Michael Grubich, President of LAK Group and adjunct faculty at Marquette's College of Business Administration, brought a practitioner's lens on talent strategy and organizational decisions across industries. Peter Hogaboam, Director of Technical Learning and Communications at CNA Insurance, shared live pilot architecture and the nuts-and-bolts of progressive risk assessment. Jim Orheim, Program Director for Executive Education and Professional Development at Marquette University, offered the academic institution's perspective on what enterprises expect from incoming talent. Dr. Phyllis King, Chief Innovation Strategist at Waukesha County Technical College, reframed the literacy-versus-skills distinction that anchored much of the exchange. Lorna Malloy, Strategic L&D Leader at Allstate, brought a Fortune 100 view of AI integration at scale — and the felt reality of how fast organizational mindset can shift. Interactive prompts from practitioners across the room surfaced the environmental friction, the silence masking resistance, and the governance gaps most organizations haven't named yet.

AI skills vs. AI literacy — and why the difference will cost you

"There's a difference between AI skills and AI literacy." — Phyllis King, Waukesha County Technical College

AI skills mean a person can operate a tool. AI literacy means they understand the effect, the ethics, and the responsible intent behind using it. One produces tool competency. The other produces organizational judgment.

Most enterprise training programs are building the former while assuming the latter follows automatically. It doesn't. Organizations that conflate the two are producing workforces that can run AI but can't govern it — and that gap shows up in compliance failures, inconsistent use, and adoption that stalls the moment someone asks a hard question about risk.

The practical implication is structural: foundational training needs to target literacy first, tool fluency second. Jim Orheim described Marquette's College of Business already requiring all faculty and staff to complete ethics, safety, and responsible-use training this summer — not because they're behind, but because they want to set the standard before the pressure arrives from outside.

Taking the seat — redefining the job, not just the toolset

"They took an organizational approach to it versus a tactical approach." — Michael Grubich, LAK Group

He described a client who stopped all incoming AI tools, stepped back to re-examine every role against business strategy, assessed whether the right people were in position, and then deployed. That sequence — organizational design before tool deployment — is what makes adoption stick.

The panel was direct: top-down AI messaging registers at roughly 4% on the behavior-change iceberg. The real work happens at the middle-manager and frontline-leader layer — and that's exactly where L&D and HR have the deepest reach. Waiting for C-suite clarity to cascade is not a strategy.

When an organization doesn't know how to rewrite roles, update operating norms, or establish communication agreements around AI use, those are not technology problems. They are people-leader problems. The moment to act on that is now.

The unspoken realities — resistance, the grid, and what's coming

The sharpest exchange in the room didn't come from the panel. It came from the floor.

A practitioner asked directly: is the environmental impact of enterprise AI data centers part of anyone's governance conversation? The panel's honest answer was largely no — and one leader named it plainly, stating that right now their organization does nothing and doesn't even have a plan yet.

What followed surfaced two signals that most organizations are not yet tracking.

The first is a psychological safety gap hiding in plain sight. Change management practitioners in the room described a dynamic that's easy to miss: people are reading the organizational climate and choosing not to surface real concerns — including environmental and ethical reservations — because they anticipate those concerns being framed as resistance to adoption. The silence looks like buy-in. It isn't. It's suppressed friction that will surface later, at a much higher cost.

The second signal is generational and incoming. Several practitioners noted that a wave of talent is arriving with environmental values baked in as non-negotiables — and that when those employees encounter mandatory AI adoption with no sustainability framework attached, the disconnect will be immediate and visible.

These aren't fringe concerns. They're governance gaps that organizations with mature people-and-culture infrastructure will be better positioned to absorb — and that organizations without it will be caught flat-footed by.

---

What to try next

1. Audit your training for the literacy gap. Pull your current AI training content and ask one question: does this teach people how to think about AI use, or only how to perform it? If the ethics, intent, and governance dimensions aren't explicitly present, close that gap before the next cohort goes through.
2. Design a listening mechanism specifically for suppressed resistance. Pulse surveys and skip-level conversations with a direct question about AI concerns — anonymous, low-stakes, framed as input rather than performance assessment — will surface what people aren't saying in public. That signal is more operationally useful than any adoption metric you're currently tracking.
3. Run one role-redefinition conversation before the next tool deployment. Before any new AI capability goes live on your team, gather the relevant people leaders and ask: if this tool does what it promises, what changes about this role? What stays? What needs to be renegotiated — including expectations, norms, and potentially compensation? That conversation, done once with intention, builds the muscle for doing it at scale.