Who Owns AI? The Culture Problem Hiding in Plain Sight (TPL Insights #319)

By Rob Andrews

In January 2026, the CEO of a Fortune 500 insurance company gathered his senior team to settle a single question: who owns AI? The CIO said it was technology infrastructure. The COO said agentic AI lives inside workflows. The CFO pointed out that AI was already making underwriting decisions with direct P&L impact. The CRO flagged autonomous systems as novel risk exposure. The CHRO called agents a new category of workers. The CDO reminded the room that none of it worked without her data governance. The meeting ended without resolution. A few months later, the company posted a job requisition for a Chief AI Officer.

This story, drawn from a recent Harvard Business Review piece by UC Berkeley professor Toby Stuart, is playing out in C-suites everywhere. Stuart frames it as a jurisdictional competition problem, rooted in sociologist Andrew Abbott’s landmark theory of how professions fight over turf during periods of disruption. He’s right about the diagnosis. But from where I sit, after nearly four decades of placing senior leaders and advising CEOs on organizational design, the deeper problem is a culture problem. And culture problems carry a direct and measurable cost in shareholder value.

The organizations that will win the agentic AI era are not the ones with the most sophisticated technology. They are the ones already operating under a peak-performance culture built on clear principles, including the one that matters most right now: clarity of accountability.

The Real Problem in That Room

One of the nine principles at the core of a peak performance culture is clear roles and accountability. Not general accountability. Specific, named, unambiguous accountability for specific decisions. In organizations that have built this discipline, “who decides?” is a routine question with a routine answer. Functional boundaries are explicit. When a new category of work emerges, leaders know how to assign it cleanly because they have built the muscle to do exactly that.

The insurance CEO’s inconclusive meeting is the predictable result when that principle is absent. Every executive in the room was right about their claim. The CIO, the COO, the CFO, the CRO, the CHRO, the CDO: each had a legitimate stake. The failure was not disagreement. It was the absence of a framework to resolve it. That gap does not just slow AI adoption. It consumes executive bandwidth, stalls investment decisions, creates duplicative work, and hands a compounding advantage to competitors who move faster. These are not soft costs. They show up in total shareholder return.

Stuart explains why agentic AI creates such intense friction. Previous technology disruptions, ERP systems, cloud computing, and SaaS platforms created confined debates. Agentic AI is different in kind. An AI agent that autonomously processes insurance claims simultaneously is the operational workflow, the risk decision, the financial commitment, the labor unit, and the data consumer. Whoever governs the agent, its roles, permissions, guardrails, and budget, increasingly controls what the company decides. The software is the service. That is why every executive in the room had a legitimate claim, and why resolving it requires more than a political compromise.

From Ownership to Decision Rights

Stuart’s most important contribution is a reframe; stop asking who owns AI and start asking who owns which AI-related decisions. The first question is zero-sum. It invites a land grab and guarantees that the executive with the most political capital wins rather than the one best positioned to decide. I have watched that dynamic damage far too many organizations over four decades.

The second question is specific and implementable. The COO or P&L owner should own what the agent is trying to accomplish and how success is measured. The CIO should own how agents are built, integrated, and maintained securely. The CDO should own data access rules and model governance. The CRO and General Counsel should own risk thresholds and escalation protocols. The CHRO should own how humans and agents share accountability. The CFO should own the investment thesis and ROI discipline.

The practical tool is what Stuart calls a decision and accountability map: an explicit document that identifies critical choices and assigns each to a specific executive. In the language of our nine principles, this is accountability clarity made operational. It is one thing to say, “I own the tech stack.” It is another to specify that the CIO owns foundation model selection, API integration standards, and the authority to shut down agentic inference if it interferes with system stability. That precision is what separates cultures that execute from cultures that negotiate.

Peak performance cultures have an inherent advantage here. Because decision rights are already explicit, integrating AI governance is an additive exercise. For organizations without that foundation, AI governance requires building it first. Those who do will find they have improved their entire operating model, not just their AI posture.

What the CAIO Role Is Actually For

Many companies are creating Chief AI Officer roles right now, often out of confusion rather than clarity. A CAIO hired to “own AI” will fail. AI touches everything, and no single executive can own everything. I have seen this pattern before: a C-suite title used to signal seriousness without doing the harder design work.

The CAIO that creates genuine value owns the coordination layer. That means building and maintaining the decision rights map, convening functional leaders when new use cases create new ambiguity, and identifying gaps before they become crises. It is a genuinely novel jurisdiction, and it is mission-critical as AI becomes the connective tissue of enterprise-scale organizations.

The profile we look for when advising clients on this hire is not a technologist who has wandered into strategy. It is an executive who understands how organizations make decisions, who holds the trust of a diverse C-suite, and who can operate across functional boundaries without needing to dominate them. That is a cultural leadership competency, and it is rare. The most effective CAIOs will be defined not by their AI expertise alone, but by their ability to build and maintain organizational clarity at speed.

The Compounding Advantage

The organizations that resolve the AI governance question fastest and most cleanly will compound their advantage at a rate others cannot match. Agentic AI, deployed with clear decision rights and sound governance, accelerates decision velocity, reduces coordination costs, scales operational capacity without proportional headcount growth, and concentrates human judgment where it creates the most value. Those are the inputs to extraordinary shareholder returns.

The turf war Stuart describes is a tax on performance. Every week it continues is a week of compounding disadvantage. The CEO who resolves it, not by picking a winner but by designing the system that assigns each executive the decisions they are best positioned to make, is doing the most important organizational health work available right now.

Culture is not a soft concept. It is the operating system that determines how fast you move, how cleanly you decide, and how effectively you deploy every resource, including the most transformative technology in a generation. Get the culture right, and the AI advantage follows. Get it wrong, and even the best AI investment will be absorbed by the friction of an unhealthy organization.

The map is the work. Build it.

Rob

Rob Andrews

Chairman & Chief Executive Officer

Celebrating 28 years of Executive Search, Executive Coaching & Culture Shaping Excellence

Direct: 713.489.9724/ Mobile: 713.301.6130

1800 St. James Place, Suite 510, Houston, TX, 77056

www.allenaustin.com Link to Allen Austin Overview

Link to Total Performance Leadership Overview

References

Abbott, A. (1988). The System of Professions: An Essay on the Division of Expert Labor. Chicago: University of Chicago Press.

Stuart, T. E. (2026, March 13). Who in the C-Suite Should Own AI? Harvard Business Review. https://hbr.org/2026/03/who-in-the-c-suite-should-own-ai