Editor’s note: The following is a guest post from Eric Johnson, CIO at PagerDuty.
The CIO role once centered on delivering reliable infrastructure and secure systems on a predictable timetable. But with the rapid adoption of AI agents in the workplace, the CIO’s job description is changing at its core, pushing beyond technology deployment into the more complex matter of human judgment.
When AI agents are used in autonomous workflows, they can make decisions, navigate ambiguity and act without a human in the loop. However, most CIOs are only now beginning to reckon with what happens when software can act independently.
Just 11% of CIOs and CTOs feel fully prepared for the scale of AI agent deployment expected over the next 12 months, according to IBM data. As organizations move from experimenting with AI to entrusting it with operational decisions, success will depend on clear governance, process discipline and understanding business context.
What should keep CIOs up at night isn’t the technology itself, but the potential gaps in its governance. In the same IBM study, more than three-quarters of organizations said AI adoption is outpacing its governance capabilities. When an AI agent can send an email, update a record, trigger a workflow or escalate an incident autonomously, the question evolves from “Can it do that?” to “Should it be able to — and under what conditions?”
This is a departure from the traditional CIO remit. Beyond ensuring performance, CIOs must decide the level of freedom systems have. That means defining the scenarios in which agents operate with full autonomy, when they require human review and when they should be kept out entirely.
For agents that aren’t built in-house, the framework should come from the organization itself. For example, an incident management agent might have full autonomy to triage and route an alert, but would require a human to sign off before the agent takes remedial action on a customer-facing incident or communicates externally.
The process discipline dividend
Successful agentic systems are built on something far less glamorous than sophisticated models or clever prompts. Organizations with documented and standardized workflows can deploy agents more consistently and at scale than their competitors.
In many ways, agentic AI is forcing organizations to confront operational gaps that have existed for years.
Workflows that rely on tribal knowledge and informal handoffs must now be documented clearly enough for an agent to follow. The result is better process discipline across the organization, which is a genuine operational dividend.
For CIOs, the need for greater process discipline marks another shift in responsibility. Deploying AI successfully now requires as much attention to operational processes as to technical architecture, because a poorly defined workflow will undermine even the most advanced AI deployment.
Well-documented processes give agents the context and constraints needed to make quick, reliable decisions. Where a major systems project once took 18 months with benefits that were difficult to measure, agentic initiatives can be launched and refined within a quarter, with results becoming visible almost immediately.
Understanding business context
The CIO in an agentic world is less a technology operator and more an orchestrator.
They are responsible for deciding where autonomous AI adds value and where the risks outweigh the benefits. Those decisions require a detailed understanding of business operations because the consequences of errors vary widely.
An agent making the wrong recommendation in an internal workflow may be an inconvenience. But a faulty decision during a high-severity customer incident could damage trust, disrupt operations and create significant business risk.
As agents take on routine work, each team’s responsibilities evolve rather than disappear. Employees are spending less time executing processes and more time overseeing outcomes and handling exceptions. CIOs who treat agentic AI as simply a technology rollout risk the human side becoming the bottleneck.
Change management and workforce readiness must be prioritized early on, because helping teams adapt to new ways of working is now part of the job. The best CIOs have always been business people first and technologists second. Agentic AI hasn’t changed that principle, but it has made it far more consequential.
Despite all the changes, security and operational stability remain core responsibilities. Agentic AI adds a new dimension by making CIOs accountable for how autonomous systems exercise judgment and how that autonomy should be controlled.
