The CIO's New Mandate
Evolve into the architect of an AI-enabled enterprise, or become its biggest bottleneck.
The fastest way for a CIO to lose influence in the age of agentic AI isn't to resist it. It's to own too much of it.
For decades, great CIOs earned trust by reducing risk, standardizing technology, and becoming the organization's center of expertise. Those instincts built resilient enterprises and deserve enormous credit. But agentic AI introduces a different kind of transformation, one that can't succeed if every experiment, every workflow, and every new capability must flow through IT.
No organization expects every employee to report through HR, or every financial decision to be made by Finance. Those functions establish standards, provide tools, ensure governance, and develop organizational capability, but they don't own every hiring decision or every budget choice. The same shift is beginning to happen in IT.
The organizations that will realize the greatest value from AI won't ask IT to become the enterprise's AI department. They'll ask CIOs to create the conditions for every business function to innovate safely, confidently, and at scale.
The CIO's job is no longer to build enterprise AI. It's to build an enterprise that can build with AI.
This isn't another ERP implementation
It's natural that organizations look to IT to lead AI, and just as natural that CIOs accept that responsibility. For the last two decades, transformative technologies have largely been enterprise technologies. ERP, CRM, cloud migrations, identity platforms, and data warehouses all followed the same pattern: the business identified an opportunity, and IT delivered the platform. That instinct isn't wrong, per se, but it’s being applied to a fundamentally different kind of technology.
The clearest precedent is actually security, not software. Modern cybersecurity isn't something IT "does" for the business while everyone else stays out of it. IT provides identity, tooling, monitoring, policies, and guardrails, but every employee shares responsibility for protecting the enterprise. A phishing test doesn't get delegated to the security team to pass on employees' behalf. Security became an organizational capability that IT enables, not a centralized service that IT delivers.
Agentic AI is following the same path, and for the same reason: the work is too distributed, too fast-moving, and too embedded in daily decisions for any central team to own on everyone else's behalf. This isn't another enterprise platform rollout. It's an enterprise capability, and it will succeed or fail based on how well it spreads, not how tightly it's controlled.
The economics of AI have changed
Within the MACH Alliance, one discussion comes up repeatedly. The future enterprise won't run on one intelligent agent; it will run on hundreds, eventually thousands, of specialized agents. Some will come embedded inside SaaS applications. Some will be purchased from software vendors. Some will be curated by departments. And many will be created (or continuously refined) by the employees closest to the work.
That future simply doesn't scale if every agent has to be conceived, prioritized, built, and governed by one central organization. The challenge for a modern AI-capable organization isn't building intelligent agents; it's building an intelligent organization.
Which leadership instincts need to evolve?
Many of the instincts that made CIOs successful can unintentionally slow this new model. Centralizing expertise, standardizing processes, controlling technology decisions, and reducing variation were exactly the right instincts for previous generations of enterprise technology.
But if every AI initiative begins with an intake form, waits for committee review, competes for the same specialists, and follows the cadence of traditional enterprise projects, the organization won't scale AI. It will scale waiting. Because the bottleneck isn't technology, but rather organizational design, and clarity of roles and responsibilities.
The strengths that matter more than ever
This isn't an argument for less IT, but it si an argument for different IT. The CIO's traditional strengths become even more valuable, not less. Rather than building every agent, IT creates the trusted foundation that allows thousands of agents to emerge safely across the enterprise. That includes:
- Security— Define what agents are allowed to access and do.
- Identity— Give people and agents role-based access to enterprise systems and data.
- Enterprise Architecture— Establish standards for how agents integrate across platforms.
- Data Stewardship— Curate trusted enterprise data for AI to use.
- Platform Engineering— Provide shared infrastructure, templates, and tooling for teams building agents, often department-specific.
- Vendor Management— Govern AI platforms, models, and strategic technology partners.
- Risk Management— Define policies for privacy, compliance, human oversight, and acceptable use.
Almost none of these involve building individual business agents. What they do is enable everyone else to.
This transformation can't be delegated
Perhaps the greatest risk of over-centralizing AI isn't technical, but organizational.
As enthusiasm for AI grows, a familiar pattern begins to emerge. Business leaders recognize the opportunity, but they're also busy, and many are still developing their own understanding of agentic AI. The easiest conclusion is a familiar one: "IT should take the lead." CIOs, eager to help the business succeed, often accept that responsibility. The reaction is understandable, but operating model is wrong.
The uncomfortable truth is that every executive—including and most importantly the CEO—must now become AI-native, must learn to build and deploy agents the way they already hire and coach employees. Sales leaders must rethink how selling changes when every account executive has a team of specialized agents. Finance leaders must redesign planning, forecasting, and analysis. HR must rethink talent, learning, and organizational design and Operations must reimagine workflows from the ground up.
The moment an executive team stops asking, "How will IT deploy AI?" and starts asking, "How will AI change the way my function operates?", the organization has crossed an important threshold. That's when AI stops being an IT initiative and starts becoming an enterprise transformation.
From there, the CIO's role becomes much clearer: provide the platforms, governance, education, and guardrails that allow every function to move quickly and safely. The organizations that move fastest won't be the ones with the biggest AI teams. They'll be the ones where every executive owns AI outcomes within their function, while IT enables success across the enterprise.
From Center of Excellence to Center of Enablement
Many organizations respond to AI by creating an AI Center of Excellence. That's a sensible place to start. In the early stages, organizations genuinely need a small group to evaluate technologies, establish standards, develop initial capabilities, and help the business learn. But over time, the Center of Excellence becomes an important bellwether for the organization's thinking. If the executive team continues to look to IT to own every AI initiative, the Center of Excellence gradually becomes a Center of Compliance, where business leaders bring ideas to IT, wait for approval, and rely on a small group of specialists to drive innovation. The capability never leaves the center.
The organizations that move fastest take a different path. Their Center of Excellence evolves into a Center of Enablement, measuring success not by how many AI initiatives IT delivers but by how effectively every business function can innovate safely on its own. IT provides the platforms, governance, education, and shared services, but the business owns the transformation.
The distinction is subtle but profound. One operating model concentrates expertise, the other distributes capability. One creates dependency, the other creates an organization that learns and scales. Ultimately, the state of the AI Center of Excellence tells you whether AI is still an IT initiative or whether it has become an enterprise capability.
The CIO's new mandate
The CIO who succeeds over the next decade won't be the one who builds the most agents. It will be the one who creates an organization capable of building the most agents safely, responsibly, and at scale.
That means centralizing what should be centralized: security, identity, governance, enterprise architecture, shared platforms, and trusted data. It also means deliberately distributing what should never be centralized: discovery, experimentation, workflow redesign, and the development of AI capabilities within each business function.
The measure of success can’t be how many AI initiatives IT delivers. Instead, it must be how many the business delivers because IT created the foundation that made it possible.
That's the real evolution of the CIO's role. Not from technology leader to business leader; CIOs have been business leaders for years. The shift is from being the organization's primary builder of technology to becoming its primary enabler of capability. The organizations that win won't have the largest AI departments. They'll have the greatest number of business leaders confidently transforming their own functions on a secure, governed, and shared enterprise foundation.
The CIO's job is no longer to build enterprise AI, but to build an enterprise that can build with AI.
Jason Cottrell
Orium Founder & CEO, MACH Alliance President
Jason Cottrell is the Founder & CEO of Orium, the leading digital commerce and autonomous transformation consultancy in the Americas, and the President of the MACH Alliance. He works closely with enterprise clients and strategic partners to prepare businesses for hybrid human–agent experiences. Under his leadership, Orium helps organizations modernize their technology foundations and unlock new levels of speed, scale, and intelligence in how they serve customers across channels.