Agentic ITSM Tools Are Arriving Faster Than Enterprises Can Absorb Them
Major vendors are already pushing agentic ITSM tools into production environments. Ivanti recently shipped an autonomous service desk agent that can create incidents, submit requests, and search knowledge bases without analyst intervention, freeing teams from repetitive tasks so they can focus on higher‑value work. Executives argue that traditional, ticket‑centric service desks are no longer sustainable, because they still depend on constant human intervention even when monitoring data is abundant. Research highlighted by McKinsey shows what is possible: one large organization automated up to 80% of roughly 450,000 annual tickets after moving to agent‑led resolution and was able to redeploy half of its service team while improving customer satisfaction. Yet this success story also comes with a warning: the company first redesigned workflows and customer journeys specifically for agent‑led resolution. Most enterprises have not done this redesign work, which is why their ITSM estates are poorly prepared for truly autonomous behavior.
The Plumbing Problem: Legacy Estates Block Enterprise IT Automation
Behind the hype, most enterprises are still stuck in pilot mode. McKinsey’s research suggests that about 62% of organizations are only piloting agentic capabilities, and no more than 10% are scaling agents in any given business function. The core issue is a "plumbing problem": legacy ITSM infrastructures were designed for linear, ticket‑based, human‑led workflows, not for agents that must cross systems, trigger actions via APIs, and operate under automated governance controls. Red Hat’s leadership has described the current moment as a back‑to‑basics phase where teams are rediscovering fundamentals like patching, highlighting how fragile many environments remain. In unified communications, a single call‑quality incident may span network telemetry, device health, carrier status, and ITSM records at once. If an AI service management agent cannot move reliably across that stack, the ticket falls back to a human operator—undercutting the promised gains in enterprise IT automation.
Why Monitoring and Governance Lag Behind Agentic Ambitions
Even when enterprises buy agentic ITSM tools, they often lack the ability to see what these systems are actually doing. Traditional monitoring confirms that a service is running but does not explain which decisions an agent made, what systems it changed, or why it acted in a particular way. Analysts warn that this visibility gap makes scaling risky: AI decisions are frequently opaque, yet errors can generate substantial financial, reputational, and regulatory damage. Gartner expects that by 2028, 40% of organizations deploying AI will have dedicated observability tooling, which implies that most businesses will remain under‑tooled for at least the next two years. Without proper observability, organizations cannot build trust in autonomous workflows, fine‑tune agent policies, or demonstrate compliance. As a result, many deployments stall after the proof‑of‑concept stage, and the tools are throttled back to assisted rather than fully autonomous operation.
Start with an ITSM Readiness Assessment, Not a Tool Purchase
To close the readiness gap, enterprises should run an honest ITSM readiness assessment before signing contracts for new agentic platforms. McKinsey’s work highlights four foundational prerequisites: a configuration management database accurate enough for agents to act on, actions exposed through APIs with embedded policy checks, a clear governance model that defines what agents can and cannot do, and active monitoring of inference costs and outcomes. For most IT and unified communications teams, this is not a simple checklist but a multi‑year program of work. Data quality is especially critical: if agents consume stale or inaccurate configuration data, they will generate errors and new tickets instead of resolving existing ones. Organizations should also review existing workflows, mapping where autonomous decisions would add value and where human oversight must remain. Done well, this groundwork turns agentic ITSM tools from risky experiments into controlled extensions of the service operation.
Building the Human and Process Foundations for AI Service Management
Technology is only part of the equation; people and processes must evolve too. Agents that can decide and act at scale demand robust governance frameworks, including role definitions, approval thresholds, rollback procedures, and clear accountability when an autonomous action goes wrong. Change management becomes more complex, because each new integration or workflow tweak can alter what an agent is capable of doing across the estate. Enterprises should invest in training service desk and infrastructure teams to understand how agentic systems work, how to interpret their outputs, and how to intervene safely when needed. This capability building helps shift the culture from ticket processing to supervising and optimizing AI‑driven operations. As financial pressure on IT infrastructure grows and AI workloads drive up costs, only organizations that prepare their data, processes, and teams will be able to use agentic ITSM tools to reduce operating effort instead of adding new complexity.