Why Managers, Not Just AI, Make or Break Transformation

Isabella

Welcome to the show everybody! I'm Isabella, here with Jesper. And Jesper, I need to start with a number from Gallup's State of the Global Workplace 2026 report that completely floored me. Global employee engagement is down to 20% in 2025, dropping from its 23% peak in 2022. But the real structural crack? Manager engagement fell off a cliff -- from 27% to 22% in just one year.

Jesper

Yes! Five percentage points in twelve months. That's not a dip, Isabella, that's an alarm bell for the entire organizational layer. We constantly talk about employee "trivsel" and engagement, but we're ignoring the load-bearing wall. If the managers themselves are disengaged, how on earth do we expect them to lift up their teams?

Isabella

Exactly, they're running on empty. And even here in Denmark, where we pride ourselves on our flat hierarchies and "tillidsbaseret ledelse" -- trust-based leadership -- we aren't immune. In fact, for Ledelsesugen 2026, employee wellbeing is the massive headline, and Lederweb and Væksthus for Ledelse are literally recruiting three entire municipalities right now just to run trials on how to protect and support the managers' *own* wellbeing.

Jesper

Which is highly logical when you think about it. If you look at the Gallup data, they found that basic, structured manager training actually cuts disengagement roughly in half. And here is the pragmatic key: it's about providing structured flexibility, not total autonomy. Total autonomy without structure just creates chaos and paralysis, which drives burnout. Yes, managers need a repeatable development rhythm, not just a single, annual "inspiration day" where they hear a nice talk and then go back to the same old grind.

Isabella

Yes! And speaking of hard, actionable proof -- have you seen the Sandoz study published in MIT Sloan Management Review? It's a randomized controlled trial -- an RCT -- with over 1,000 teams and 7,000 people globally. That is massive! They finally have *causal* proof on what actually builds psychological safety.

Jesper

A randomized controlled trial with 7,000 people in a corporate setting is incredibly rare. And what did the data actually show as the winning lever? It wasn't putting up posters about "safe spaces" or doing trust falls. The single biggest driver of psychological safety was managers treating team members as unique individuals. Personalization beats generalization every single time.

Isabella

Yes! They measured things like "different perspectives are valued in my team" and "I feel safe sharing feedback." And treating people as individuals is a trainable behavior! It’s about asking specific questions, understanding their unique working styles, and checking in. And if you want a free, immediate culture audit on this, look at the HBR research on interruptions. Who gets cut off in a meeting before finishing their thought? That is a live psychological safety signal right there.

Jesper

Yes, absolutely. If you want a quick diagnostic tool for your next Danish client on Monday, just sit in their team meeting and count the interruptions. Who interrupts whom? If the manager is constantly cutting people off, or letting others do it, you have your answer. It's a simple, low-cost way to make "tillidsbaseret ledelse" highly concrete and measurable.

Isabella

It really is. Now, let’s bridge this over to why major change initiatives fall apart. A March 2026 Harvard Business Review article re-analyzed that infamous McKinsey stat -- you know, that 70% of organizational transformations fail -- and they pinned the blame squarely on leaders who lack basic people skills. Specifically, leaders who misread silence as buy-in and dismiss genuine, constructive concerns as mere "complaining."

Jesper

This is such a critical point, and it connects so deeply to Christian Ørsted's work here in Denmark, especially in his book *Fatale forandringer*. He argues that what leaders call "resistance to change" is almost always a completely rational reaction to poor, uncollaborative leadership. When leaders treat silence as consent, they are designing their own failure.

Isabella

Yes! HBR offers four fixes for this: diagnose the skill gap without making it personal; build interpersonal skills through continuous repetition, not just a one-off seminar; redesign the system to compensate for leadership blind spots; and honestly, know when you have to develop a leader versus when you simply have to replace them.

Jesper

Yes, you have to build the habit. If you don't actively hunt for dissent before a rollout, you're flying blind. Easy, immediate unanimity in a meeting is actually a massive warning sign, not a success. Now, let's talk about where this lack of alignment is playing out right now in real-time: AI adoption. The CBS AI Survey from April 2026, led by Professor Torsten Ringberg and published in Djøfbladet, surveyed 4,281 Danes. And the gap is staggering.

Isabella

Oh, the 38-point gap! Leaders are at 70.1% active AI use at work, while the average employee is down at 41.8%. That is a massive digital divide within the same organization! And to make it worse, 84% of top executives admitted they have no real AI strategy. So we have highly enthusiastic leaders racing ahead, using the tools, but with absolutely no strategic direction for their teams.

Jesper

It is the ultimate recipe for what Dennis Nørmark calls "pseudoarbejde" -- pseudo-work. If you give people powerful generative AI tools without a strategy, they don't necessarily do more valuable work; they just produce *more* slide decks, *more* reports, and *more* emails that nobody actually needs to read. We are manufacturing digital busywork at an unprecedented scale.

Isabella

It's so true. It's like Morten Münster's behavioral design framework in *Jytte vender tilbage* -- if you want good AI adoption, you can't just lecture people or buy them licenses. You have to design the environment so that using AI for high-value work is the path of least resistance. Otherwise, they'll just use it to automate the useless tasks.

Jesper

Yes! If you want your team to be a high-performing "Superteam" -- referencing Ron Friedman's book -- you have to focus on how the team actually learns and adapts together. Friedman shows that the teams that learn the fastest are the ones that win. So before buying a single AI license, write a one-page "AI intent" document. Define exactly what value you want to create, and audit the workflow for AI-enabled pseudo-work.

Isabella

I love that. A one-page intent. It's so simple but it forces clarity. So, to wrap this up for our consultants heading out to clients on Monday: First, remember the manager is the load-bearing wall -- invest in their repeatable development and structured flexibility first. Second, psychological safety is a trainable behavior -- treat people as unique individuals and audit those meeting interruptions. And third, silence is not consent -- actively hunt for dissent, and don't let AI become a factory for pseudo-work.

Jesper

Yes! So let's leave our listeners with one final, pragmatic question to chew on: Is your team actually high-performing and improving -- or are you just leading on trust in theory, while tracking the wrong metrics in practice?

Isabella

Oof, a great question to take into the week. Thanks for listening, everyone, and we'll catch you next time!

Jesper

Yes, thank you, and take care of each other out there.