Essay · May 2026

Capital has a CFO. AI has a CAIO. Decisions have no one.

A position firmly stated and loosely held on the next C-suite role.

Last month Meta named its first Chief AI Officer. FactSet did the same. They’re the latest in a fifteen-year pattern: every time the C-suite expands, a new role gets installed to govern a resource that crossed some threshold of strategic importance.

The next one is already here, and almost no one is naming it. Here’s a position firmly stated and loosely held: by 2030, most Fortune 500 companies will install a Chief Decision Officer, accountable for how decisions get made across the organization by humans and increasingly by agents. Feel free to convince me otherwise.

I spent years on the consulting side of large transformations, sitting inside PMOs at BCG. The model works. When you bring focused human expertise, structured analytics, and decision discipline to a Fortune 500 trying to navigate change, you get better decisions and better outcomes. That’s why companies have been buying this work for half a century.

What’s changing is the unit economics. Historically, the capacity to run a continuous, enterprise-wide decision governance function was prohibitively expensive, so companies bought it episodically, project by project. AI shifts that calculation. The PMO function becomes something that can run continuously, embedded in the org, at a fraction of the historical cost.

When the function becomes continuous, someone has to own it permanently. I think that someone is a Chief Decision Officer, accountable to the CEO, governing how decisions get made by humans and agents across the organization. The job is to constantly build, optimize, and improve the system of record and decision rights for how humans and agents make choices, then monitor the outcomes and feed those signals back to improve the next generation of decisions.

The pattern is already in motion

The C-suite has expanded the same way for decades. CFOs emerged when capital became too distributed for CEOs to govern alone. CIOs in the 1990s when information systems became strategic. Chief Information Security Officers when security became existential. Chief Data Officers when data became an asset class. The pattern is consistent: a resource becomes strategically critical, distributed across functions, lacking a single owner, and compounding in risk. Then a role appears.

Chief AI Officers are the most recent. In 2023, fewer than 11% of Fortune 500 companies had one. As of early 2026, that number is 26%. By the end of this year, it’s projected to hit 40%. In just the last 60 days, FactSet appointed its first Chief AI Officer, and Meta named Alexandr Wang to the same role following its $14 billion talent acquisition. Each new role expansion has standardized faster than the last.

Why decisions, why now

Three forces collided.

First, the discipline matured. Lorien Pratt coined “Decision Intelligence” in 2008. Gartner declared it “transformational” in its 2025 AI Hype Cycle. Google ran an internal Chief Decision Scientist function for five years, training 20,000 employees. More on that below.

Second, the economic case sharpened. Bain found top-quintile decision-effective companies deliver shareholder returns six percentage points higher than peers, with 95% confidence correlation across countries and industries. McKinsey’s 2019 survey found only 20% of executives say their organization excels at decision making, and 72% in a 2017 study said bad strategic decisions were as frequent as good ones or the prevailing norm.

Third, AI agents are forcing the issue. Gartner projects 50% of business decisions will be agent-augmented by 2027. Sequoia’s Julien Bek argues the next trillion-dollar AI company will be a software company “masquerading as a services firm,” because the prize is selling outcomes, not tools. Foundation Capital makes the parallel argument from the infrastructure side: the next trillion-dollar platforms will be systems of record for decisions.

We have mastered the art of recording what happened. We are still terrible at capturing why.

Jaya Gupta & Ashu Garg, Foundation Capital, December 2025

Here’s the part I find most interesting. AI doesn’t just make decision governance cheaper. It makes a previously-impossible coordination problem tractable. No single human could hold the full context across marketing, operations, finance, product, and supply chain at once. AI can. The CDO is the role that turns that capability into governance.

Isn’t this just the Chief AI Officer?

It’s the most common conflation, and worth disentangling. The Chief AI Officer governs the AI capability stack: model selection, infrastructure, talent, deployment, vendors. Their job is to make AI work inside the organization. The Chief Decision Officer would govern something different: how decisions actually get made across the organization, by humans, by agents, or hybrid. Decision telemetry, decision rights, audit, agent governance.

Put another way: a CDO is to decisions what a CFO is to capital. The CFO doesn’t run the banks. They govern how capital is allocated, audited, reported, and improved. In practice, great CAIOs and CDOs would work closely. But they’re not the same role.

The closest existing analogs aren’t the CAIO either. The Chief Strategy Officer is closest for what gets decided. The Chief Transformation Officer is closest for howdecisions get made during major change, though project-bounded rather than continuous. The Chief Decision Scientist is the most direct named precedent. None of them add up to the role I’m describing.

The Google counter, honestly

The strongest argument against my thesis is the one Google itself is making. From 2018 to 2023, Cassie Kozyrkov ran the role I’m describing as Google’s first Chief Decision Scientist. She left in 2023, and Google didn’t replace her. They didn’t drop the function either. They embedded it across Responsible AI, People + AI, and human-centered design teams. The capability got absorbed rather than centralized.

That works for Google. Their business is cross-functional data infrastructure, and their bench depth lets them embed what most companies can’t. A consumer goods company, an insurance carrier, a hospital system is structurally functional. Cross-functional decisions fall through the cracks because no executive has standing to own them. For companies like that, centralization is how the role gets onto the org chart in the first place. Embedding may be the endgame, but you don’t get there without first naming someone responsible.

The open questions

This is where I’d most welcome other reads. Four questions I’m sitting with.

I think the function is coming. I’m less certain about the shape. Feel free to convince me otherwise.