The opening keynote of the Gartner IT Infrastructure, Operations & Cloud Strategies Conference Autumn S. and Paul Delory landed at a moment where many IT leaders already feel the ground shifting under their feet. Cost pressure has not eased. Talent remains scarce. Business expectations around speed have increased again. And now AI agents enter the picture, not as an experiment, but as something executives expect to see delivering outcomes.
Early in the keynote, the tone was set with a line that drew laughter and discomfort at the same time: “I am so glad that we could all get together one last time before we all get replaced by AI.” The humor matters because it reflects a real tension. IT leaders are being told AI is inevitable, yet they are also expected to keep systems stable, secure, and affordable as they adopt it.
I believe we can all agree that AI agents will transform IT operations, but not in the simplistic way many headlines suggest.
AI Is the Answer, but the Question Still Matters
One of the most important exchanges in the keynote came early. Gartner asked a simple but uncomfortable question. If AI is the answer, then what exactly is the problem we are solving?
CIOs are under strong pressure from the business to deploy AI quickly. Cost reduction dominates those conversations. Gartner cited that cost-cutting is the top priority for a majority of CIOs heading into the next planning cycle, with AI widely seen as the mechanism to achieve it.
That framing is risky because when AI is positioned primarily as a cost-cutting tool, IT organizations are pushed toward shortcuts, automation is rushed, controls are deferred, and governance is seen as friction rather than an enabler. We have seen this pattern before with cloud adoption.
From a CxO perspective, this is the first reframing that needs to happen. AI agents should not be introduced on a slide because they are not backed by evidence. They should be introduced because they change how work gets done, how decisions are supported, and how scale is achieved. Cost benefits follow only when those foundations are in place.
AI Agents Are at the Peak of Hype
Gartner placed AI agents at the top of the hype cycle, and based on my LinkedIn feed, I can't imagine how much more hyped they could get. That matters because hype distorts priorities. When something is overhyped, it attracts executive attention, budget, and expectations faster than organizations can adapt their operating models.
But while AI agents dominate conversations, most day-to-day operational challenges still sit further along the maturity curve. Infrastructure automation, DevOps practices, continuous delivery, and service management are not new, but they are still unevenly implemented. These are the areas that drive reliability, resilience, and predictability.
For many organizations, AI agents are being layered on top of operational foundations that are not ready. That creates a fragile environment. Agents depend on data quality, access models, stable interfaces, and clear ownership. Without those, they exacerbate existing problems rather than solve them.
In customer conversations, this shows up quickly. Teams experiment with Copilot or automation agents, but outputs vary widely. Some agents are helpful, others surface outdated content, overshare information, or fail in unpredictable ways. The issue is rarely the model; it is the operational context around it.
Doing Nothing Is Not an Option
One of the clearest statements in the keynote was also one of the most pragmatic: “Doing nothing is not an option.” Gartner acknowledged that skepticism exists inside IT leadership. Many leaders have seen waves of technology come and go. Yet the business expectation around AI is already set.
This creates a narrow path for IT operations leaders. They must move forward, but they must do so deliberately. The keynote framed I&O teams as the group that turns ideas into action. That responsibility has not changed; what has changed is the pace and the ambiguity.
AI agents are not delivered as a single product with a fixed scope; they evolve continuously. New capabilities appear with short lead times, defaults change, and models improve, requiring a different operational mindset. Traditional change control processes struggle to keep up.
From a CxO angle, this means expectations need to be reset internally. Success with AI agents is not about a single rollout; it is about building adaptive operational capacity. That includes faster evaluation cycles, clearer ownership, and stronger feedback loops between business and IT.
The Risk of Becoming the Bottleneck
One of the most striking warnings in the keynote focused on relevance. Gartner stated that if IT operations do not adapt, they risk becoming the bottleneck. When that happens, the business will find another way, through shadow IT, outsourcing, or decentralized tooling.
This is not a theoretical risk; we see it already with AI. Employees experiment with external tools, teams connect third-party agents, and business units automate workflows without involving central IT. The intention is speed, the outcome is fragmentation.
Gartner clearly described the long-term consequences. Once IT is viewed only as a cost center that maintains the status quo, it loses influence, budget, and talent. That is the real risk for CxOs. Not AI replacing people, but relevance eroding quietly.
To avoid this, IT operations need to move from reactive control to proactive enablement. That does not mean removing guardrails. It means designing guardrails that scale.
From Supporting Technology to Shaping Outcomes
The keynote challenged a long-standing self-image within IT operations. For years, I&O teams have been positioned as support functions. Keep the lights on. Maintain uptime. Control cost. AI agents change that equation.
Agents do not just execute tasks; they influence decisions. They shape how information is accessed. They operate at a scale and speed that humans cannot match. That makes them part of the operating model, not just part of the toolset.
Gartner posed a hypothetical that deserves serious consideration. What if I&O were out in front, building today what the business will need tomorrow? That requires a shift in mindset. IT operations becomes a value-driving function, not just a stabilizing one.
In practice, this means closer alignment with business outcomes. It means understanding which processes benefit from agentic automation and which do not. It means measuring success beyond uptime by evaluating decision quality, cycle time, and risk reduction.
AI Agents and the Reality of Operational Complexity
One of the most honest moments in the keynote was when Gartner listed the implicit expectations placed on IT operations. Move faster. Innovate. Deploy new technology. Cut costs. Reduce headcount. Maintain security. Keep existing infrastructure running.
This tension is not new, but AI agents intensify it. Agents introduce new dependencies, rely on permissions, consume data, incur variable costs, and require monitoring.
In our customer environments, we often see early agent deployments create hidden operational work. Someone needs to track which agents exist. Someone needs to understand what data they access. Someone needs to respond when outputs cause issues. Without structure, this work becomes manual and reactive.
This is where the conversation about governance needs to mature.
Governance as an Operational Capability
The keynote did not frame governance as a compliance exercise. Instead, it positioned governance as an operational necessity. That aligns closely with what we see in practice.
Traditional governance models assume relatively static systems. AI agents are not static. They change behavior as models evolve and as underlying data changes. Governing them requires continuous visibility and adaptive controls.
From a CxO perspective, governance should answer practical questions. Which agents exist? Who owns them? What data do they touch? How often are they used? What do they cost? What risk do they introduce?
Without those answers, scale is impossible. With them, AI agents become manageable components of the digital workforce.
Learning From the Cloud Era
Looking back to the cloud journey of the past two decades, any organization moved too slowly at first, then too quickly, and paid the price in sprawl, cost overruns, and security gaps. This is why Rencore was built: to help organizations of all sizes regain and maintain control of their Microsoft 365 environments.
AI agents risk following the same pattern. Early caution turns into unchecked expansion; retrofitting controls later is painful and expensive.
The lesson for CxOs is clear: invest early in operational foundations. That includes data hygiene, access models, lifecycle management, and cost transparency. These are not blockers to innovation; they are prerequisites.
Customer Experience From the Field
In my conversations with large enterprises, a consistent pattern emerges. Teams that succeed with AI agents treat them as part of operations, not as experiments. They define ownership, they set expectations, they measure usage and impact.
One global organization described its first agent rollout as a wake-up call. The technology worked. The outputs were impressive. But within weeks, they realized they had no inventory of agents, no clarity on cost drivers, and no way to explain unexpected behavior to auditors.
Their second phase focused less on new agents and more on operational structure. Once that was in place, adoption accelerated again, with far less friction.
The Role of the CIO and the Executive Team
AI agents are not an IT-only concern; they reshape how organizations operate, making them a leadership issue.
- CIOs need air cover to say no to reckless speed and yes to sustainable progress.
- CFOs need transparency into cost models that are no longer fixed.
- CISOs need visibility into non-human actors that access sensitive data.
- COOs need confidence that automation supports, rather than destabilizes, core processes.
This requires shared language and shared metrics. AI agents should be discussed in the same forums as workforce planning, risk management, and operational excellence.
Reframing Success in the Era of AI Agents
This leaves us with a forward-looking message. I&O can be proactive partners. They can enable new initiatives and make technology easier for the business.
That vision is achievable, but only if success is defined correctly. Success is not the number of agents deployed. It is not the speed of rollout. It is the ability to scale AI safely, predictably, and with clear value. AI and agent governance can be the solution, not the bottleneck, when innovation is rapid.
For CxOs, the question to ask is simple:
Are we building an operating model where AI agents increase leverage without increasing chaos?
Answering that question honestly is the first step toward transforming IT operations in a meaningful way.
