CoE Creative Leadership 2026 Annual Symposium

We are at the cusp of a transformation that will make organizations, work, and markets unrecognizable. Yet most companies are currently automating processes that were designed to fit the division of human labor, not AI agents.

If you were building your organization from the ground up today — no legacy systems, no inherited org chart, no “this is how we’ve always done it” — with AI as the founding principle, how would you go about it? What opportunities would you prioritize?

This is the essence of the challenge facing most companies today: Becoming AI-First.

Two keynotes frame the intellectual stakes. Two panels of business leaders bring the operational reality. A select group of CEOs and C-level executives, in an interactive format designed for peer conversation.

Keynote 1: Marshall Van Alstyne

“Deepfakes and Deep Decisions: Using Markets to Clear Misinformation”

AI generates misinformation at industrial scale. Hallucinations are mathematically inevitable. Deepfakes are cheap to produce and expensive to detect. The information environment on which every organization depends is degrading.

The problem has a precise economic structure. Misinformation is an externality: like pollution, it imposes costs on others that the producer does not bear. Externalities cause market failures. Market failures normally call for regulation. But regulating speech — even false speech — is constitutionally constrained and politically intractable. Is there an alternative? A market-based mechanism that makes truth cheap and lying expensive, without requiring a central authority to judge what is true? One has been designed, experimentally tested with thousands of participants, and validated at scale. This keynote explains how it works.

Keynote 2: Shane Greenstein

“Commercial Breakthroughs and the AI Gold Rush”

The five largest technology companies will spend between $600 and $690 billion on AI infrastructure this year. Capital deployment at this scale has happened before, in railroads, electrification, fiber optics. The pattern is remarkably consistent: the investment creates lasting infrastructure, even if individual participants lose.

What separates winners from the rest is not the technology. It is “co-invention”: the organizational redesign required to translate a general-purpose technology into specific business value. New workflows, new roles, new decision processes, work that doesn’t necessarily register as technology investment but determines whether the investment pays off. The dot-com crash destroyed hundreds of companies. The infrastructure they financed became indispensable, but only for the organizations that redesigned how they operated around it. The same question now faces every company deploying AI. This keynote explains what history teaches about where the value actually accrues.

Panel 1: “Work Reimagined: Humans, Agents, and the New Division of Labor”

McKinsey now describes its workforce as 2/3 humans and 1/3 agents. Every organization will need to find its own ratio and it will be different for every team, function, and level. But the workforce mix question is only the beginning.

New competencies are needed — but which ones, and how do you develop them at scale, fast enough? Performance itself also needs rethinking, as a way of understanding how people learn and help each other to create value from AI. These questions need urgent responses to defend against the thinning of entry-level roles and to sustain the pipeline that cultivates senior expertise.

And then the deeper challenge. AI is formally a tool, but it reasons, communicates, and exercises judgment. People have no vocabulary for something that is simultaneously a subordinate instrument and a superior performer. It is an identity crisis for knowledge workers and leaders alike.

Panel 2: “Building AI-First Organizations: Strategy, Operations, and Competitive Advantage”

Most companies have deployed AI by automating existing work, while the longer term gains are likely to be in creating new value. How does strategy balance the two approaches?

Part of the answer is maturity. Companies at the experimentation stage need different leadership and strategy than those attempting to scale, and different again from those redesigning their business models around AI. But what does it take to move from one stage to the next, and how do you dismantle structures that have been working well?

The time horizon compounds the difficulty. Organizational transformation has a payback period measured in years. Boards and markets think in quarters. And even when efficiency gains materialize, the Jevons paradox applies: the more productively you use AI, the more you use it, generating new complexity, new cost, and new demands that could even absorb the gains entirely. What, then, does a credible AI strategy actually look like?