ITSM in a World of AI: D.O.A. or A.O.K.?

---

A deep look at why IT Service Management is evolving—not dying—in an AI-driven world, and why governance and standards matter more than ever.

ITSM in a World of AI

D.O.A. or A.O.K.?

Over the last 40 years, I had a lot of alphabet soup. Starting with COBOL, ISAM, VSAM, IMS, CICS. JCL and scores of more letter combonations that followed.

Every few years a new set of technology letters comes along and eats the existing ones. What about ITSM, information technology service management. Will the latest set of letters, (and everything else for that matter) be eaten by AI?

I think hardly. At least not ITSM. In fact, ITSM will be needed even more in the ever-growing complexities of AI.

Is ITSM Dying?

Is ITSM dying? No! In fact, it is evolving and growing more necessary in order to support artificial intelligence and all its flavors.

Rather than blowing away ITSM, AI is pushing it into a new era: adaptive, predictive, and experience driven service management.

Traditional ITSM focused on stability and control, constant delivery of services to ensure IT shops kept running smoothly. Many of these services will now have AI behind them to more effectively and efficiently provide the those services ITSM has always provided.

AI-Enabled ITSM

AI enabled ITSM focuses on anticipation and continuous improvement. Instead of waiting for incidents, AI will help to be proactive in providing its services.

In an AI Co-pilot query, the response was:

Instead of static SLAs, organizations move toward dynamic experience level agreements (XLAs) informed by real time telemetry. Instead of siloed processes, AI encourages integrated value streams where data flows freely across development, operations, and business teams.

That’s AI about AI.

But this interpretation misunderstands what ITSM actually is. ITSM has never been about tickets, forms, or rigid process flows. At its core, ITSM is a management philosophy: a structured way of ensuring that technology services are reliable, valuable, and aligned with business needs.

AI can automate tasks within ITSM, but it cannot replace the governance, accountability, and strategic decision making that ITSM provides.

Governance, Standards, and Risk

In fact, AI makes ITSM more essential. As organizations adopt increasingly complex systems—hybrid clouds, microservices, distributed architectures—the need for disciplined service management grows.

AI may detect an anomaly, but determining whether to act on it, how to prioritize it, and how it affects business outcomes still requires human judgment guided by ITSM principles.

AI may generate a change plan, but the organization still needs a governance framework to assess risk, approve deployment, and ensure compliance.

More important than the above quote from a Co-pilot query, our ventures and adventures into AI need not just frameworks but adherence to strict standards to assure interoperability of agent components.

My students know that my favorite use case on the important adherence to standards is the NASA's Mars Climate Orbiter (1999), which disintegrated in the atmosphere because of a fundamental unit conversion error (metric vs. imperial) in navigation software.

Lack of standards in agentic AI components can cause disasters like that.

ITSM Managing the Complexity of AI Itself

It’s not just that ITSM can take advantage of AI to provide the services it does. Even more so it works the other way around with the complexity of AI systems, ITSM is needed to manage the complexities of AI itself.

Take take a recent guide for Agentic AI that I received on Linked In. It discussed the allocation layer, orchestration layer, tool layer, memory layer, there were seven layers in all.

Each layer doesn’t have cake and icing, but it’s own components.

Managing the Complexity of Agentic AI

Agentic AI agents can be built from modular components created by different organizations, and in fact that appears to be the dominant architectural pattern in the industry.

Components allow you to mix and match vendors, swap out components without rebuilding the whole system, add new capabilities over time, maintain control over sensitive data.

Just take a look at an AI architectural diagram and you’ll quickly see how complex and how many components could be included in an agent.

Here’s an example; an agent that contains a ServiceNow API (from ServiceNow), CMDB lookup service (internal), Risk scoring model (built by a third party vendor).

How is an organizations going to keep track of all of the components?

Although it was before the AI explosion, I often think of and mention in my lectures the Solar Winds disaster.

Organizations that knew they had Solar Winds and or knew solar winds was a component in another piece of software they were running were able to isolate their vulnerabilities quite rapidly.

The Bottom Line

ITSM is not dying. It is the organizations that fail to move quickly to advance their ITSM to meet the challenges of AI, may die, or crash on Mars.

Rather than killing ITSM, AI is moving it into a new age.

Traditional ITSM functions which may be taken over in some part by AI but will still be necessary for IT shops to function.

The day I am writing this is the NASA Artemis rocket is being moved to the historic Cape Canaveral 39b launch pad. Hope their ITSM is doing well, and the standards have been impeccably enforced.