Q&A | October 21, 2026

How DXC OASIS and DXC Agentic SOC are bridging the divide between cybersecurity and IT operations

Within the systems that power modern business, a gap has long existed between IT infrastructure operations and cybersecurity. This is an issue that must be addressed to secure long-term operational success.

Historically, the IT Ops team, led by a CIO, has managed the day-to-day maintenance and episodic upgrades to ensure that mission-critical systems do not fail.

In parallel, the cybersecurity team has built systems to check for vulnerabilities and secure an enterprise from cyberthreats.

The two teams existed in the same technology ecosystem, but they often operated in silos — independently running and protecting the enterprise.

The reality of today’s AI-enabled cyberattacks demands a different approach. With attacks operating at machine speed, attack surfaces expanding, and safety, reliability and governance requirements only increasing in importance, now is the time to raise an important question: Does the naturally occurring silo between IT Ops and cybersecurity teams expose an enterprise to heightened risk levels?

Together, DXC's President, Global Infrastructure Services Dan Gray and Vice President, Global Chief Information Security Officer Mike Baker asked and answered this question with a new organizational construct: Human operators work alongside AI agents to ensure that DXC’s IT estate is protected and operates without failure. In theory, the two teams coming together already tracks. But in practice, only DXC OASIS, the company’s intelligent orchestration platform, makes this construct possible. DXC OASIS unifies AI security capabilities, including DXC Agentic SOC (DXC’s agent-driven Security Operations Center), into one observability and orchestration interface. With DXC OASIS as the unified front, capabilities, including incident response and patching, can be achieved in real time. With DXC Agentic SOC as a key agentic cyber-offering, cyberthreats are identified and addressed at machine speed with specialized cybersecurity agents operating alongside operational agents — like the Linux agents that are core to DXC OASIS.

In the following conversation, Dan and Mike discuss the operational need for greater integration and the organizational impacts of this shift.

Q: Historically, IT and cybersecurity have largely operated independently. How is the need for operations to react at machine speed changing that paradigm?

Mike Baker: The paradigm has changed because the threat has changed. Traditionally, an IT operations shop has operated independently, and a cybersecurity shop has operated almost as a check and balance to that. But given that threat actors are leveraging AI to operate at machine speed, the typical delay between those teams, when they operate in silos, will not and cannot work anymore with these adversarial threats hitting you not within weeks, days, or hours, but within minutes or seconds.

The seam that existed — the delay, the lead time to get things done across IT Ops and cybersecurity teams — can be almost completely eliminated when we bring DXC OASIS into the picture.

Dan Gray: The threat profile that we face now through AI is incompatible with a purely human effort. Every organization traditionally deals with lists of open vulnerabilities. And it's up to the CIO and CISO to work together to prioritize those. The best leaders work very closely and have combined teams, as Mike and I have. But we're at a point where you can't just rely on that upper percentile of human leaders and analysts. We need to shift certain identification and analysis tasks from human operators to agentic capabilities to operate at machine speed.

Q: How is provisioning and identity management evolving as organizations move toward machine-speed AI operations?

Dan Gray: It's all about adding value to businesses. You’d think that it doesn’t matter if humans or agents operate business processes, since both are working in the background and follow a specific set of predefined tasks. But once we start bringing agents into the fold, you have what we're calling background or ambient agents. These agents are running continuously, separated from the core identity of the human orchestrating them. This is straightforward when it's done one-to-one — when my agent is almost a digital twin of me and my permissions. But imagine the explosion in complexity when I have a team or a fleet of agents that can do new and different things, apart from the role and tasks that I have.

Mike Baker: As we embrace agentic technologies, the agents who do the work are going to decouple from the individuals. To Dan's point, that creates a new problem: How do you manage these digital agentic identities at scale, where companies could conceivably quadruple in size without adding human employees? What we've been doing at DXC is — number one — making sure our existing non-human identity landscape is sound and its security and governance hygiene tightly managed.

And — number two — we’re teaming with great partners, who, along with our own product teams, are defining what the identities of tomorrow look like. Identities that are spun up very quickly, that are assigned to agents that can perform work, that are just in time; and that are deprovisioned very quickly. These identities can be created natively with guardrails from position descriptions or other mechanisms.

This is all future focused. But the key thing here at DXC is that, just like with our agentic SOC, we're looking around the corner and rapidly testing solutions as the pace of change increases. 

Q: How do agentic orchestration systems like DXC OASIS handle vulnerability mitigation, patching, and operational recovery at machine speed while ensuring safety and compliance?

Mike Baker: A requirement Dan and I are working on right now is our need to identify exploitable vulnerabilities at machine speed and have them move through an agent harness, like DXC OASIS, that can patch those vulnerabilities. We're working on different levels of patching, which include applications as well as an entire operating system. Both of which involve either a human-in-the-loop or human-on-the-loop approach — in addition to other guardrails — to govern the workflow and to ensure that we properly test processes the way we would use them in practice. Thus, we can apply those patches in a quick, responsible manner.

Dan Gray: AI allows us to do things very, very quickly. However, you need a harness to execute and apply that to specific environments. A regulated or enterprise environment will have very specific and unique change controls regarding things like patch sequencing or burn-in time. That is the context that we integrate into DXC OASIS and use as the basis for process orchestration. 

Q: DXC OASIS was designed with flexibility in mind to operate across many platforms. How important is an open tool architecture for enterprise operations?

Dan Gray: It's absolutely essential. I don't think there would be another way of creating the reference architecture without that open tool layer below it. Having an open architecture that allows us to integrate and interact with multiple tools is one of the key design characteristics and considerations of DXC OASIS.

It's important to note that every product now either has, or is adding, AI capabilities. DXC OASIS does not try to compete with those capabilities. Instead, it integrates them. It's the layer of work across the top.

Q: What is the evolving role of human talent, as organizations adopt agentic AI and systems like DXC OASIS and the agentic SOC?

Mike Baker: At DXC, we believe that AI amplifies human potential. In our SOC, entry-level cybersecurity analysts learn the new tools, become AI-native and can accelerate their career pathway, bringing these technologies to our customers. Basic tasks, like working with reports and spreadsheets, are completed by AI, freeing humans to focus on things that aren't mundane.

Dan Gray: There’s a definite change in the roles that humans play versus AI agents. In roles that run things on high-frequency, that are highly execution-oriented, that are traditionally available 24/7, agents work really well. This allows the human team members to really elevate their skills and roles within the team. What we are seeing is that those humans — who have a real open mind to new technology, a strong expertise and an aptitude for systems thinking — are the true superpower in the AI era. They encapsulate the experience and knowledge of the top 1% of our engineering community and deploy that across all our customers.

Q: Is there a point in the future where IT and cybersecurity become a single, integrated team?

Mike Baker: More importantly, it’ll be a team of agents working together with an integrated team of humans orchestrating those agents across the board. We see a trend where security professionals are taking on infrastructure operational roles at an increasing rate. They’ll continue to merge. And at some point, they'll work in concert with agents.

Dan Gray: It's a multiplayer mode between agents and humans to solve tasks. In version two of DXC OASIS, which is right around the corner, we’ve built-in capabilities for this multiplayer mode, where multiple humans can interact together in a Teams or Slack channel with multiple agents.

Q: What are you hearing from customers? What is keeping them up at night?

Dan Gray: It feels like right now everyone's being kept up at night by the fear of one thing: a long-chain, zero-day vulnerability that nobody has a solution to. And if we had a solution, how long would it take to apply it?

Mike Baker: I'm hearing from some of our customers about either a lack of boardroom support or that they don't have a solution to the regulatory requirements. So, they don't know where to start. They see there's certainly value to be unlocked, but they just need some help to get out of the batter's box and get to rounding the bases.