Article | April 27, 2026

How AI is rewriting the business of running IT

By Chris Drumgoole, President, Global Infrastructure Services, DXC Technology

The new competitive advantage is not where work gets done. It is how well organizations can orchestrate specialized AI agents alongside their people to deliver outcomes at a speed and scale that was not previously possible. Labor arbitrage as the primary competitive lever in managed IT services has run its course. AI is here, and the new competitive differentiator is agent arbitrage: how well you orchestrate intelligent, purpose-built agents alongside human expertise to deliver outcomes at machine speed. The companies and providers who have already made that shift are pulling ahead. The ones still debating it are falling behind.

And yet, most of the conversation in our industry right now is outward-focused, on customer-facing innovation, on announcing the next AI investment, on what's new and external. That's important work, but there is an equally significant opportunity that is getting less attention: using AI to fundamentally change the business of running IT itself. The new and the existing. The exciting and the operational. Both matter.

This is not a story about the next chatbot or AI tool bolted onto a helpdesk. It's a structural shift in how enterprises run IT across three core domains: cybersecurity, infrastructure and workplace. And the gap between the leaders and the laggards is widening every day.

The conversation boards are actually having

Boards are pushing CEOs and CIOs hard right now. They see competitors announcing new investments and innovations daily. They see cost pressure and productivity pressure converging at the same time. And they're asking direct questions:

Can we run this business better with AI?

Can we redirect our people toward higher-value work?

Can we move faster?

These aren't theoretical questions anymore. At DXC, we've taken service resolution times from 3 hours to 30 seconds. We've reduced security investigation times by more than 70%. We've shifted entire teams from reactive ticket processing to proactive engineering and defense.

This isn't an incremental improvement. It's a before-and-after moment for how IT operates.

Cybersecurity: You can't fight machine speed with human speed

Security is being transformed the fastest because it must be.

Attackers are already using AI. They're automating reconnaissance. They're generating adaptive phishing campaigns at scale. They're probing environments around the clock at machine speed. The old model of humans staring at screens triaging an endless alert queue is now indefensible.

If you aren't fighting AI with AI, you're already behind.

At DXC, we deployed agentic AI into our own security operations, using ourselves as Customer Zero, before asking any customer to trust it. AI agents now triage alerts, correlate threats, and escalate intelligently. That shift reduced investigation times by 77% and freed our security experts to work on higher-value work.

What happens when you remove the alert queue from a security team? They stop reacting and start hunting. They build defenses with intention. And because they're no longer buried in routine tasks, they invest time in understanding each customer's environment, anticipating threats before they materialize rather than scrambling after them.

The threat landscape is growing more complex. Pairing intelligent automation with disciplined human oversight is more than an efficiency play: it's how you stay ahead.

Infrastructure: From ticket factories to something that actually scales

Infrastructure quietly became something no one planned for: a bifurcated maze of tools, dashboards and ticket queues where the primary unit of work is still a ticket or incident moving between humans.

We layered DevOps on top of it and expanded into cloud. But strip it back and most enterprise environments are still fundamentally run the same way they were 20 years ago. A person sees a problem, creates a ticket, assigns it; someone resolves it, closes it, moves to the next one. Repeat, indefinitely, at whatever scale your headcount allows.

That model doesn't scale enough. And it's colliding head-on with a moment when the demands on infrastructure have never been higher.

Resilience, reliability and compliance requirements are more complex and more consequential than at any point in the history of enterprise IT. As some recent high-profile outages have shown, an IT failure is no longer just a bad day. It's a business stoppage, with regulatory, reputational and financial consequences that land immediately and visibly. The infrastructure running your business has never mattered more, and the model being used to run it has never been more inadequate.

When a task that once took hours can be completed in seconds, the economics of IT change. And, importantly, this changes what your best technical people spend their time on.

When you remove the mundane as a constraint, your smartest engineers stop clearing queues and start solving business problems. Infrastructure stops being a cost center and becomes a competitive lever. The CIO running it earns a seat at the table not just as an operator, but as a driver of business performance.

Xponential, DXC's AI orchestration blueprint, sits at the heart of how we do this. Built with embedded governance, compliance and observability, it ensures AI and agentic frameworks operate reliably at enterprise scale. Not a product. A discipline. The approach behind every solution we deploy.

Getting here requires two things to happen simultaneously. Leaders need platforms that give engineers real-time visibility and control across the entire IT estate. And those engineers need to be reskilled to orchestrate AI agents rather than respond to alerts. That's a people investment as much as a technology one.

Workplace: IT should feel like a flagship retail experience

Enterprise IT support systems were not originally optimized to ensure employees got help — they were optimized to meet a faceless Service Level Agreement (SLA) at the lowest cost. The industry called the metric a "deflection rate" and considered a high score a success. The entire architecture — tiered support models and self-service portals nobody actually used — was designed to minimize human contact, not maximize resolution.

Walk into a premium consumer electronics store and they don't give you a ticket number when you have a tech issue. They don't tell you to check the FAQ. They fix your problem while you wait. That expectation that technology just works without training, without a manual, without a three-day wait, is now so deeply ingrained that employees bring it through the door every morning. And then they hit your service desk.

AI gives us the opportunity to close that gap. Not by adding another chatbot layer on top of a broken process, but by actually redesigning how employee support works. Automated resolution where it makes sense. Immediate human expertise where it doesn't exist. Meeting employees in the tools they already use, in the format they actually want, because a junior engineer and a senior finance director may not want the same support experience, and now we don't have to pretend they do.

When DXC implemented AI-powered automation for Textron, we cut service desk tickets by 20% while proactively resolving network issues for tens of thousands of employees across 23 countries. Routine problems handled autonomously. IT teams focused on what actually requires human judgment.

This is what we mean by Human+: AI handles the repetitive, predictable work, so people can focus on the meaningful, complex and creative work that actually moves the business forward. The goal isn't fewer interactions. It's better outcomes.

A word on what AI actually is in this context

I've noticed a consistent pattern in executive conversations about this. Senior leaders often think AI can do far more, far faster, than it realistically can do right now. Engineers can give you a detailed list of reasons it won't work. Both are wrong in different ways.

This isn't about deploying one large model and hoping for transformation. It's not about ChatGPT running your data center, or Claude rewriting 50 years of COBOL code overnight.

The real capability comes from agentic frameworks: small, purpose-built models orchestrated together to execute specific tasks with speed and reliability. Think less "one big brain" and more "a highly specialized, tireless digital workforce." When deployed correctly, with governance and human oversight built in from the start, these agents don't just answer questions, they also provide guidance. They act. They remediate. They optimize. They learn within defined guardrails.

That's how you actually run IT with AI.

Do more. Move faster. Spend less. All three.

Managed services have always operated under a fundamental tension: customers want higher quality and lower cost, and historically, you could optimize one or the other, not both.

AI changes that equation, but only if you're willing to fundamentally reshape how IT operates rather than just automate what you already have.

That requires speed. It requires governance. It requires reskilling teams to orchestrate AI rather than process tickets. And it requires proof, which is why we first prove everything internally before taking it to customers. We compete every day with organizations that believe they can do this themselves. Some can, but many underestimate the complexity of running AI-enabled IT at scale across environments they didn't build from scratch.

This is not a science experiment. It's an operational discipline.

For decades, DXC has been running mission-critical systems for the world's leading enterprises, including airlines, governments, financial services leaders and manufacturers. We understand not just the infrastructure, but the outcomes those systems support. AI doesn't change that responsibility. It raises the bar.

The real opportunity

The question isn't whether AI will reshape IT operations. It already has.

The real question is what you do with that fact.

CIOs who use this moment to simply automate what they have will run a slightly more efficient version of the same cost center. CIOs who use it to fundamentally redesign how IT supports the business will find themselves in a different conversation entirely, one about competitive advantage, business velocity and strategic direction rather than ticket volumes and resolution times.

When you eliminate the 3-hour ticket. When reactive security becomes proactive defense. When your engineers spend their time on strategic work instead of clearing queues. You run IT better. You run the business better. And everyone in that boardroom knows the difference.

The seat at the table isn't given to the person who manages IT most cheaply. It's earned by the leader who makes the business faster, safer and more capable than it was before.

That's the conversation we should have. And there's never been a better moment to start.

About the author