The latest bright and shiny object in the IT service management space is artificial intelligence.
But, despite the benefits of AI in an ITSM implementation, operations teams need to carefully consider whether they need AI capabilities, as well as whether they can sustain and support AI-based tools and processes.
While ITSM theory has long promoted an "ideation to retirement" approach to manage IT services and products, many ITSM implementations have focused only on the operational aspects of service management.
In many instances, an organization has a service desk and established incident management and request fulfillment processes. Some organizations have defined a basic change management process that funnels requests through a central approval body. While these approaches might be beneficial, they are far from a comprehensive and holistic ITSM strategy.
As new approaches to ITSM -- such as DevOps, VeriSM, lean IT and shift left -- have emerged, the frameworks of ITIL, ISO/IEC 20000 and COBIT have undergone significant revisions. IT asset management and business relationship management have become part of the ITSM mix.
Technologies like cloud computing, IoT, containers, microservices, big data, automation and, of course, AI in ITSM have made practitioners rethink how they should structure ITSM processes. In addition, new business drivers, such as enhanced customer and employee experience, have also caused ITSM to evolve.
IT teams need to balance the responsiveness that the digital economy demands with organizational stability. They're looking to exploit ITSM to achieve both of these goals.
AI-enabled ITSM market
AI technologies can enhance or augment an organization's approach to IT product and service management. Most current AI-enabled ITSM systems focus on the following technical areas, with the first two -- service desk capabilities and IT operations -- being the most common:
- Service desk. Offload tedious and repetitive tasks, such as routine issue resolution or request fulfillment.
- IT operations. Enable event management through event filtering and correlation
- Incident management. Categorize incidents based on defined criteria, as well as automatically route tickets or remediate some incidents, without human intervention.
- Request fulfillment. Automatically fulfill end-user requests, often through a self-service model, including any required approvals.
- Problem management. Use large data sets to detect patterns and correlate IT incidents.
AI-enabled ITSM platforms typically fall into one of three categories: reactive, proactive or predictive.
Reactive tools augment the service desk, primarily through end-user self-help and deflecting contacts. They perform actions in response to a stimulus, such as an alert, or human interaction. Some examples of reactive AI in ITSM tools include chatbots and virtual assistants.
Editor's note: With extensive research into the ITSM tool market, TechTarget editors have focused this series of articles on vendors with considerable market presence and that offer ITSM tools with AI functionalities that can be classified as responsive, proactive, predictive and autonomous. Our research included Gartner, Forrester Research and TechTarget surveys.
Proactive products detect a condition or pending situation and take action to mitigate it before service interruption occurs. Automation, robotic process automation (RPA) and process orchestration tools are typical entrants in this category.
Predictive offerings forecast future performance, demand or IT needs, based on analysis of historical and current data. Tools in this category might produce new data sets from gathered data sets to make these predictions. Examples are ITSM tools that use machine learning and make big data relevant and actionable.
Conceptually, there could be a fourth category of AI-enabled ITSM tools: autonomous. Autonomous tools can make value judgments and take actions independently of their initial set of algorithms. These platforms can interpret and apply complex concepts, such as risk/reward or the ethical implications of potential actions, to ITSM. At this time, however, there are no commercial autonomous tools in the ITSM market.
Do you need AI-enabled ITSM?
Here are some common cases that might prompt an organization to adopt an AI-enabled ITSM platform:
- Augment (limited) staff resources. This move doesn't mean an organization will no longer need actual IT staff. Instead, the company can refocus IT staff on knowledge development and the design of processes, workflows and algorithms.
- Eliminate IT operational friction or siloed activities. AI-directed tools promote standardized and consistent approaches, and can highlight bottlenecks and anomalies.
- Centralize multiple, disparate sources of service management data. Most IT organizations collect data that relates to IT services from sources other than an ITSM tool, such as tools for network monitoring, application performance management and IoT deployments.
Requirements for AI-enabled ITSM adoption
While AI-based IT tools offer many benefits and meet a range of use cases, it's best not to rush adoption.
First, ensure your organization has well-defined ITSM processes that consistently deliver expected results. An IT organization should have clearly articulated processes for frequent tasks involved with incident management and request fulfillment, as this uniform approach is what enables automation. Admins should periodically review these processes to ensure ongoing relevancy and appropriateness.
The value of IT services to the business should also be clear, and the IT organization should value knowledge from a methodology perspective, rather than a strictly technological one.
In addition, AI adoption for ITSM requires a well-defined and well-managed configuration management database (CMDB). While AI sciences, such as machine learning, have the capability to "learn," this learning is based upon clean and thorough input data. If the CMDB does not appropriately represent how components underpin services, AI-based ITSM efforts will struggle. IT teams must define and periodically review the development of a data strategy, in general, that addresses topics like governance, integrity and retention to ensure that the use of AI provides long-term benefit. Additionally, make sure knowledge bases are AI-ready -- meaning, they're up to date and relevant to the organization.
New and expanded AI-enabled ITSM tools debut almost daily. But before an organization commits to a tool, it needs to fully understand its own IT and business needs.
First, determine the business case. What's the reason to adopt an AI-enabled ITSM tool, and what are the anticipated benefits and return on investment? While the use of these tools will improve efficiency and speed, cost-cutting and job elimination are the wrong drivers for implementation. Effective use of AI will shift the use of human resources from lost time on tedious, repetitive work to productive work that involves creativity, value judgment and strategic thinking.
Evaluate current competencies, and ensure the skills necessary to exploit and maintain AI capabilities are available in-house. Having in-house skills such as process design, value stream mapping and knowledge article development will help organizations take full advantage of a platform’s AI capabilities.
Assess how accepting users, and IT staff, will be of AI-enabled self-help. If members of the IT organization fear the loss of their jobs, carefully communicate the intended use of AI-enabled ITSM tools.
AI isn’t a silver bullet
Like good ITSM, the implementation and use of AI within ITSM will not be a one-and-done thing. The use of AI will not magically fix poor process designs or address any silo mentalities that exist within an organization. Be clear on the initial problems you want to solve and the desired outcomes before you make an investment in AI-enabled ITSM tools.