31 Executives Name AI and Judgment as Leadership's New Test

By Olivia Hart

Thirty-one members of the Fast Company Impact Council say artificial intelligence has moved leadership from debate to redesign. Their answers to a single question — what trend is reshaping leadership today — cluster around one theme: AI changes what managers manage, not whether managers exist.

The sharpest claim comes from Michael Weening, CEO of Calix. He argues AI adds a sixth job to the manager's classic five. "AI doesn't eliminate managers. It changes what we manage," Weening said. He tested the shift himself, spending a month reteaching himself to code for the first time since he left computer science for a business degree in 1987, building an application by routing tasks across several AI models.

What does AI actually change for managers?

For decades, Weening writes, a manager's job came down to five things: setting goals, aligning work, coaching people, clearing obstacles, and judging results. The sixth, in his framing, is mapping and improving the workflows a team executes.

Blake Brannon of OneTrust pushes the logic further. He says the job is shifting from making decisions to articulating the judgment behind them, because executives can now embed their thinking directly into the tools their teams and agents use. His verdict is blunt: "AI hasn't made leaders less necessary. It has made vague leadership much harder to hide."

That claim deserves scrutiny. Nothing in the 31 answers offers evidence that embedding context into tools improves outcomes; the argument rests on assertion. What the council does document, though, is a consistent worry about leaders who set AI strategy without understanding daily work.

Thomas Scott of Wrike calls this out directly. Too many executives, he writes, form AI strategies "based on what they see in the market without understanding how work happens today." His takeaway: leaders need firsthand insight into where AI makes a meaningful difference and where human judgment still matters.

In plain words: the council wants bosses closer to the workflow before they buy the tools.

Who manages the agents?

Two answers sketch a structural change to careers. Tigran Sloyan, CEO of CodeSignal, says high performers may skip the people-manager path entirely and instead manage a team of AI agents as individual contributors. Organizations, he writes, will need to rethink how they recognize impact and build career paths for those contributors.

Tony Grimminck of Scribd, Inc. describes the same shift from the other side. With agentic AI, professionals now direct "armies of digital agents" that execute tasks and write code, but humans must stay in the loop. His conclusion: "AI doesn't eliminate the need for human judgment; it significantly raises the bar for it."

Hanneke Faber, Logitech's CEO, adds a limit to the enthusiasm. AI can generate answers all day, she writes, "but it does not engage human teams to act, nor can (or should) it make tough decisions for us." Her bet: decisive leaders who motivate through co-creation — ownership, belonging, control, agency — will make actual business progress.

The caveat writes itself: none of these executives quantifies the productivity or cost effects they describe.

Where is the skepticism inside the council?

One member supplies it. Manuel Freire-Garabal of the Gioya Health Education Institution warns of a "prompt-and-send" culture, where people use AI-generated material without reviewing, understanding, or validating it. The consequences surface later, he writes, when staff cannot explain the subject in presentations or implementation.

His prescription: strengthen human thinking and use AI as a tool to support knowledge, responsibility, and expertise — not as a substitute.

Darren Person of Cengage gives the worry a cost frame. Skipping the learning curve on employee AI experimentation, he writes, "can lead to inefficient usage and token overages." The opportunity, in his telling, is not universal AI adoption but identifying what works and using those lessons to rethink how the organization operates.

Beyond AI: what else did the council flag?

The 31 answers split into several non-AI camps:

• Distributed ownership. Susan Watts of SPACECRAFT runs her organization as "a pie, not a pyramid," where great ideas should emerge from any slice.
• Insourcing. Yuping Huang of Quantum Computing Inc. says owning the company's fabs, rather than relying on someone else's capacity, reflects a broader lesson: know what is critical and be willing to own it.
• Politics as an operating condition. Howard Franklin of Ohio River South argues "we stay out of politics" has stopped being a strategy, because rules that held for decades now change in a news cycle — and must be managed in statehouses and county commissions, not just Washington.
• Values-based promotion. David Angelo of David&Goliath advocates "casting from values": choosing leaders for what they believe and embody, not only what they can do.
• Constant principles. Andrea Montecchi of Oliver Wight Americas draws the line at integrity, accountability, judgment, courage, and respect for people: "The environment changes. The methods change. The principles do not."

ServiceNow's chief people officer, Jacqui Canney, frames the organizational challenge most ambitiously: redesign the work first, then reimagine the workforce, then make sure "the workbench is ready" — platform, data, and agents. Her warning is that leaders treat AI's democratized intelligence as a technology problem when the real challenge is organizational: developing judgment at speed without eroding trust.

Adam L'Italien of Liberty Mutual Insurance compresses the whole survey into one sentence: capabilities that define the frontier one quarter become commonplace the next, so staying ahead "stops being something you can delegate and becomes the job itself." The fastest learners, he writes, will win.

That is a promise, not a finding — the council offers no data on which leaders are actually winning. What the 31 answers establish is where executive attention has moved: away from whether to adopt AI and toward who directs the agents, who validates their output, and who still makes the judgment calls.