The Boston Consulting Group released its 2026 AI Radar last week, and one finding cut through the rest. Nearly three-quarters of CEOs now describe themselves as their company's chief AI decision-maker. Average enterprise AI spend is set to roughly double in 2026, moving from 0.8 percent of revenue to 1.7 percent (BCG, 2026).

This is not a delegation moment.

It is an authorship moment.

For most of the last decade, the AI conversation lived a layer below the CEO. CIOs led pilots. CTOs led architecture. Strategy teams led use-case selection. The board got an update once a quarter. The CEO weighed in at decision points but did not own the trajectory.

That model is closing. The data BCG put on the table says the chief executive is now sitting in the chair. The question is which version of that role they intend to play.

This is what closes the May series, and it is the strategic frontier of the next decade.

This Month’s Arc

This edition closes the May series on Human-Centered Leadership in an AI-Driven World:

• How a 75 percent CEO ownership figure from BCG is reframing the AI conversation as an authorship problem, not a delegation problem

• Why operator-mode and steward-mode chief executives will produce identical short-term results and radically different decade-long ones

• What Deloitte's 2026 governance research and PwC's CEO Survey reveal about the gap between executive ambition and institutional capacity

• Why a Chief AI Officer appointment does not solve the steward problem (and what does)

• Three concrete practices the leaders building durable AI-era institutions are using now

THE AUTHORSHIP MOMENT
What the BCG Data Just Made Visible

BCG's 2026 AI Radar surveyed CEOs across more than 30 markets and found that nearly three out of four now identify themselves as the chief AI decision-maker in their organization. The average enterprise AI budget will rise from 0.8 percent of revenue to roughly 1.7 percent over the next twelve months. CEOs remain bullish on investment and optimistic on returns (BCG, 2026).

The companion BCG research published the same week framed the implication directly. AI has made work reinvention a CEO mandate, not a CIO project (BCG, 2026).

This shift is the headline. The substance is what the shift makes possible, and what it makes dangerous. It makes possible a generation of chief executives who author the conditions of AI deployment in their organizations, not just approve them.

It makes dangerous a generation of chief executives who deploy faster than they can govern, and discover only later that the productivity gains came with structural commitments they cannot reverse.

A recent CIO publication highlighted the gap. Seventy percent of marketing chiefs cite AI leadership as their top 2026 goal, while only thirty percent believe they have the infrastructure to deliver on it (CIO, 2026). MIT Sloan's 2026 action-items research for AI decision-makers describes the same gap at the executive level. Ambition is high. The institutional capacity to govern that ambition is not (MIT Sloan, 2026). This is the operator's ceiling. The organization can deploy faster than it can govern. The CEO becomes a faster executor of a strategy that no one has authored yet.

THE OPERATOR
What Most CEOs Are Becoming By Default

An operator is a chief executive who treats AI as a system to be deployed. The questions they bring to the room are operational. Which workflows. Which vendors. What return on investment. What productivity gain.

The cadence is quarterly. The success metric is throughput.

The Foxconn and BCG case study cited in the AI Radar describes an agent ecosystem that automates 80 percent of a category of decision-making and unlocks roughly $800 million in value. Fujitsu's AI agents in supply chain reduced warehousing costs by $15 million and halved staffing needs (BCG, 2026). The capabilities are real. The momentum is real.

But the operator frame quietly compounds risk that does not appear in the dashboard.

PwC's 29th Global CEO Survey, published this year, found that despite high investment, only 12 percent of CEOs say their AI deployments have measurably improved profitability so far (PwC, 2026). Deloitte's 2026 governance research adds a sharper finding. Ninety-nine percent of executives believe boards should be using AI in their oversight role. Only thirty-five percent of directors say their boards actually are (Deloitte, 2026).

That gap between belief and practice is the operator's signature. The organization moves; the governance does not. By the time the gap is visible in performance, the structural commitments are years deep.

THE STEWARD
What the Next Decade Will Require

A steward is a chief executive who treats AI as something the organization must learn to govern, not just to deploy.

The questions they bring to the room are different. Who is accountable when this agent acts. What standard will we hold our partners to. Which decisions will we not delegate to a model, on principle, even when we technically could. What does it mean for our customers, our regulators, our workforce, if we get this wrong.

The cadence is generational. The success metric is durability.

Novo Nordisk's April 2026 partnership with OpenAI is worth studying carefully through this lens. The company announced full integration of OpenAI models across drug discovery, clinical development, manufacturing, supply chain, and corporate functions, with scaled deployment by late 2026 (Drug Discovery World, 2026; Fierce Pharma, 2026). What separates the announcement from many similar ones is the structural posture. Novo Nordisk explicitly framed data governance, human-in-the-loop oversight, and workforce upskilling as core deliverables of the partnership, not after-thoughts (BioPharm International, 2026).

The decision was reported as bold. The more interesting question is who at Novo Nordisk authored the conditions of that integration. The boundaries. The audit cadence.

The regulatory framing. The standards. The steward authors those conditions deliberately. The operator does not.

FUTURE OF CTOS
The Chief AI Officer Question

A common response to the authorship problem is to appoint a Chief AI Officer. Roughly 38 percent of enterprises have done so or have an equivalent role in place (BCG, 2026; MIT Sloan, 2026).

Naming the role is the first step. It is not the answer.

The 2026 AI and Data Leadership Executive Benchmark Survey found that CAIOs report to the CEO 43 percent of the time, to the CTO or CIO 35 percent of the time, and to the COO 12 percent of the time (MIT Sloan, 2026). At JPMorgan, the AI-focused executive sits on a 14-person operating committee reporting directly to Jamie Dimon (MIT Sloan, 2026). At many other enterprises, the role sits two layers down from the chief executive and reports through a technology function that has not yet developed the regulatory, ethical, and institutional muscle the role requires.

PwC's CAIO research published this year frames the issue cleanly. The CAIO role is becoming central, but its authority depends entirely on whether the chief executive has authored the governance conditions the role is meant to execute (PwC, 2026).

Where those conditions are authored at the top, the CAIO role becomes a force multiplier. Where they are not, the CAIO becomes the senior person tasked with managing the consequences of decisions made without governance design.

What The Best Chief Executives Are Doing Now

Three practices separate the steward-mode CEOs I see from the operator-mode default.

First, they author the standard before they buy the tool. They name explicitly which decisions in their organization will remain under human authority regardless of AI capability, which will be augmented, and which will be delegated. Procurement follows the standard. The standard does not follow procurement.

Second, they design the audit cadence into the rollout, not after. The Agentic AI Institute reported earlier this month that 72 percent of enterprises have agentic AI in production, but a 60 percent governance gap remains. The steward-mode chief executive closes that gap before the first agent is live. The operator closes it after the first incident.

Third, they invest in the institutional literacy that AI governance actually requires. Deloitte's most recent governance research found that 40 percent of board members and senior executives say AI has caused them to reconsider their boards' composition (Deloitte, 2026). The steward-mode chief executive treats board literacy on AI as a strategic capability to develop, not a procurement question to delegate.

These are not soft practices. They are the institutional architecture of the next decade of leadership.

SHARE YOUR THOUGHTS
We value your feedback!

Your thoughts and opinions help us improve our newsletter. Please take a moment to let us know what you think.