(01)
Radical imagination
Sure, AI improves what organizations already do, but it also dismantles the rules those organizations were built around. Leaders must think in terms of massive sea changes rather than incremental improvements.
We've seen this play out with BMW, where leaders chose to use machine learning to predict crash performance, reducing reliance on late-stage physical prototypes and drastically altering safety engineering. One leader in our network said it best: "Leaders are starting to really get in there, and it sparks their imagination. When I experience AI firsthand, it enables me to connect the dots differently."
(02)
Permission to be wrong
In an environment where AI capabilities aren't specified in advance, the leaders gaining ground are those who have turned experimentation from an occasional activity into an operating system.
Consider Intuit, which built a proprietary generative AI platform, GenOS, which enables product teams to move new AI ideas into live customer experiments in days rather than months. Teams across the company have tested hundreds of GenAI use cases through the platform, and leadership revised internal rules that had limited AI test rollouts to just ten customers, expanding the threshold to 1,000—a deliberate move to accelerate learning rather than contain risk.
(03)
Thinking modularly
"When line-of-business owners lead and IT is riding shotgun as an enabler—that's where we see organizations move fastest," one leader told us. As AI unbundles jobs into tasks and enables work to move in parallel, the old model of planned, sequential execution breaks down.
Duke Health's implementation of AI-powered operational management is a study in what should come next: Leaders centralized real-time data across the hospital system, embedded machine learning into daily decision-making, and redistributed authority so that coordination became continuous rather than episodic, yielding a 66% reduction in bed request-to-assignment time. This breakthrough is attributed to leadership deliberately redesigning how work flows.
(04)
Discerning proactively
A tidbit from one leader captures the new discipline required in the AI era: "Everything I do now starts with, 'What's the role of AI? What's the role of humans?'" AI systems can now draft communications, recommend actions, and operate through agents. But who decides when to trust them, when to intervene, and where to draw the line between humans and machines? That responsibility rests with leaders.
Salesforce deployed its AI agent (Agentforce) on its own help site. Early on, the team instructed the agent not to discuss competitors—but it overcorrected, refusing to help a customer integrate Microsoft Teams. Leadership replaced rigid rules with a simpler guiding principle: act in the customer's best interest. Human handoffs were designed into the system from the start, and escalation to a person is treated not as a failure, but as the right outcome when nuance is required.
