The Agentic Advantage for ITSM Chapter 4

Use Case Selection & Prioritization

When evaluating agentic AI for ITSM it’s easy for organizations to start by asking, “What can we automate?” The better question is: "What should we automate first?"

Here’s why: Most enterprises have hundreds of service desk processes, thousands of workflows, and countless opportunities for automation. Attempting to automate everything at once almost always leads to stalled pilots, unclear ROI, and organizational resistance.

The most successful organizations take a different approach. They identify a small number of high-impact use cases, prove value quickly, establish governance and trust, and then expand over time. They start with economics. The goal is to identify the largest sources of SaaS tax and eliminate them first.

Following the money

Traditional ITSM teams often prioritize use cases based on visibility. Organizations focus on the loudest complaints, the most complex incidents, or the processes receiving the most executive attention. While understandable, this approach frequently overlooks the largest automation opportunities.

The most effective strategy is to follow the money. Look for requests that are:

• high volume
• highly repetitive
• well understood
• policy driven
• labor intensive
• expensive to fulfill manually

These characteristics are important because they directly influence the economics of service delivery. Every repetitive request requiring human intervention consumes valuable resources. The same principle applies to software access requests, device provisioning, VPN troubleshooting, onboarding activities, and other routine service desk functions. Through this lens, the best automation opportunities become obvious.

The most effective pilot use cases

Across industries and enterprise environments, a consistent set of use cases repeatedly emerges as the best starting point for autonomous service delivery.

Password resets

Password resets remain one of the highest-volume support requests in most organizations. The process is typically well understood, governed by established policies, and relatively low risk.

Most importantly, it is a request employees expect to be resolved immediately. Password resets represent an ideal opportunity for auto-resolution because the desired outcome is clear and measurable. The employee needs access. The system provides access. No ticket required.

Account lockouts

Account lockouts share many of the same characteristics. They are common, repetitive, and disruptive to employee productivity. Auto-resolving account lockouts delivers immediate value to both employees and IT teams.

Software access requests

Software access requests often involve a surprising amount of coordination. Approvals, entitlement checks, license assignments, and provisioning activities can create significant delays. These processes are excellent candidates for agentic execution because they typically follow established policies and involve predictable decision paths.

The opportunity is not simply approving requests faster. It is completing the entire fulfillment process automatically.

Software provisioning

Once access is approved, software still needs to be delivered. Historically, this required human intervention. Today, orchestration across identity systems, endpoint management tools, and software repositories can enable end-to-end fulfillment. The result is a dramatically improved employee experience.

VPN and connectivity support

VPN issues represent another common source of service desk demand. While some connectivity problems require investigation, many follow recognizable patterns that can be diagnosed and resolved automatically. The ability to identify, troubleshoot, and remediate common connectivity issues can significantly reduce support volumes.

Distribution lists and collaboration tools

Adding employees to distribution groups, collaboration channels, shared drives, and communication platforms is often highly repetitive work. These requests rarely require complex decision-making but generate substantial administrative effort. They are ideal candidates for autonomous fulfillment.

Device requests and replacements

Whether provisioning a new laptop, replacing a damaged device, or assigning equipment to new employees, hardware-related requests often involve predictable workflows. Automating these processes can reduce administrative overhead while accelerating delivery times.

Employee onboarding and offboarding

Few IT processes are more visible than onboarding and offboarding. These workflows typically require actions across multiple systems from identity platforms, HR and security systems to endpoint management tools and collaboration platforms.

Because the steps are generally well defined, onboarding and offboarding represent powerful examples of end-to-end agentic execution. Employees receive everything they need faster. Organizations improve security and compliance, and IT teams reduce manual effort.

The 80/20 opportunity

One of the most important concepts in autonomous service delivery is the 80/20 model. Most organizations discover that approximately 80% of service desk demand consists of repeatable, process-driven work such as password resets, account unlocks, software requests, etc.

These activities follow established policies and require limited human judgment. The remaining 20% is different. These requests involve complex incidents such as risk-sensitive decisions, exception handling, and strategic judgement.

The goal of agentic AI is not eliminating the need for humans. The goal is ensuring human-in-the-loop expertise is applied where it creates the greatest value. When organizations automate the repetitive 80%, service desk teams gain more time to focus on the 20% that genuinely requires their knowledge and experience.

This shift changes the role of the service desk. Teams spend less time resetting passwords and routing requests and more time solving meaningful problems.

Prioritizing for auto-resolution, not automation

As organizations evaluate opportunities, it is important to distinguish between automation and auto-resolution. Automation may improve a workflow. Auto-resolution completes an outcome. This distinction should influence prioritization decisions. The best use cases are not necessarily the easiest to automate. They are the ones capable of delivering complete outcomes without human intervention.

A chatbot that explains how to request software may reduce ticket volume. A system that automatically provisions software changes the economics of service delivery. One provides information. The other eliminates work. Organizations should prioritize accordingly.

Building a roadmap beyond IT

Many organizations begin their journey within IT. The service desk typically contains large volumes of repetitive work and well-documented processes. However, the same principles often apply across the enterprise. Once organizations establish success in IT, they frequently expand into other areas of the business.

The underlying objective remains the same: Identify repetitive work. Apply agentic execution. Deliver outcomes. Reduce operational burden. Over time, autonomous service delivery becomes an enterprise capability rather than an IT initiative. The goal is not to automate everything; it’s to eliminate the most expensive sources of routine work, prove value quickly, and build a foundation for broader autonomous service delivery across the enterprise.