AI in Leadership: What 30 Senior Leaders Learned About Performance and Culture

No PR or press. Just a pressing question on the table: as AI changes how work gets done, which assumptions about leadership and people strategy still hold — and what needs to change?

The answer, it turned out, was even more interesting than expected. Here's what we found out.

Why AI adoption is slow in some organizations

We opened the summit by asking the room to place themselves on an AI maturity curve — Experimenting, Integrating, Embedding, Differentiating. Not a single organization placed itself at Differentiating. Almost everyone landed between stages one and two.

That can’t be chalked up to the technology, it’s ready and available. What isn’t in place, though, are the systems around it: the performance models, the talent structures, the leadership behaviors — all designed for a world where value was measured in execution and output. That world hasn’t disappeared, it’s just no longer enough.

AI embedding is stalling because performance systems haven't yet caught up, because no one's agreed on who owns the AI agenda internally, and because leaders are advocating for transformation they haven't personally been through.

The organizations moving fastest? They're redesigning those systems. Not just deploying more tools.

The role of culture and trust in AI transformation

Each session explored how people leaders are navigating this moment of change. One theme that ran through all of them: culture is the infrastructure needed in times of resilience.

Hillary Champion, VP of Talent Management at Netflix, made this case directly. Netflix takes an atypical approach to performance: no ratings, no performance improvement plans, no traditional reviews. Instead, there's a foundational belief in trusting people completely — and a simple test of managerial intent: would we genuinely fight to keep this person?

What makes that possible is high transparency, high trust, and real autonomy in decision-making. The implication for AI transformation is the same: organizations that have built genuine trust cultures can move faster, because their people have the judgment and psychological safety to experiment without waiting for permission.

Trust cultures also require something specific from leaders. You have to be on the journey yourself. Leaders who advocate for transformation from a distance — without being visible, active participants in it — signal, however unintentionally, that change is for everyone else. That signal travels fast, and it's one of the most reliable ways to stall adoption before it starts.

Vidya Krishnan, Chief Learning and Belonging Officer at TD SYNNEX, brought the data that made this concrete. Her organization deployed a digital learning platform with 18,000 courses alongside a coaching program. The platform got a 30% take-up rate. Coaching got 95%. The difference wasn't content — it was relationship. People want to learn from and with other people. That's what trust enables, and what process alone can't replicate.

Paddy Hull, Global Head of Talent, Culture and Leadership at Heineken, added a dimension the room hadn't fully accounted for. The loneliness epidemic is real and measurable: 25% of the global population is experiencing deep aloneness, with mortality rates now comparable to those linked to obesity. The workplace is one of the few institutions with the consistent reach to counter this. Organizations that build for belonging are making a commercial decision as much as a human one.

How AI is changing leadership skills and behaviors

Dr. Haesun Moon closed Day 1 by tracing the etymology of the word 'intelligence' back to the Latin inter-legere — to gather between. Intelligence has never been a solo act. It has always lived in the space between people.

AI distributes information faster and further than any individual ever could. The human contribution isn't to compete with that. It's to direct it, question it, and make meaning from it together. The most consequential thing a leader does is decide what questions are worth asking.

EZRA's research identifies four qualities that consistently separate the organizations advancing on the AI maturity curve from those that stall: Curiosity. Discernment. Humility. Connection.

Here's what struck us in Rome: those four words surfaced independently across table discussions on both days — without prompting. We didn't put them on a slide and ask people to agree. They arrived at them through their own experience of what is and isn't working.

The leadership standard, as one attendee put it, has shifted from skill-based to character-based. How a leader shows up — their willingness to admit uncertainty, their capacity to ask rather than direct — is now the primary signal of AI readiness.

Key leadership lessons for AI transformation

Five things worth holding onto — whether you were in Rome or not.

1. Stop measuring AI adoption by usage rates. The organizations getting the most from AI aren't the ones with the highest adoption scores. They're the ones where people are bringing judgment, curiosity, and discernment to every interaction with it.

2. Redesign the systems, not just the tools. Performance models were built for a different distribution of work. Deploying more AI into a system designed for the old one doesn't close the gap.

3. Be on the journey personally. Endorsing transformation isn't the same as being part of it. Visible learning, not just visible support, is what moves organizations.

4. Trust is the infrastructure. High-trust cultures don't need as much process. They also perform better through disruption. That's not a coincidence.

5. Make development travel. Individual leaders are growing through coaching. That growth isn't yet moving fast enough into teams, culture, and systems. That's the next frontier.