The ITIL 4 body of service management best practices was launched in early 2019, with the main guidance delivered in early 2020. While this might not seem that long ago in “IT service management (ITSM) years,” it is in “artificial intelligence (AI) years.” When the main elements of guidance were written, AI use in ITSM was still a future opportunity for many organizations. And as such, it was highly unlikely to be included in any detail. After all, ITIL consolidates industry good practices for the benefit of all.
This blog looks at how more than the ITIL 4 guidance is needed for your IT organization to fully benefit from AI, with the IT service desk and related capabilities (or management practices) a good place to demonstrate the gap between ITSM AI use in 2019 and 2025.
Using IT service desk management practices as an example
The scope of IT service desk practices will differ across organizations. For example, some might consider change management (or change enablement in ITIL 4) an IT service desk capability, while others won’t. Some will undertake proactive problem management as a separate capability. In contrast, others will use problem management techniques only in their major incident management practices.
No matter what’s included or not, these five management practices are still good examples of how ITIL 4 references and helps with AI adoption:
1. Service desk
2. Incident management
3. Service request management
4. Change enablement
5. Problem management
1. Service desk
The latest service desk practice guide (accessed January 2025) does not mention artificial intelligence or AI. It does, however, have three machine-learning references in the following sections:
1. General information – “When using machine learning, ensure a high quality of data and algorithms.”
2. Information and technology – “Leverage machine learning capabilities: They can significantly optimize both self-help experience for users and query categorization activity for the service desk agents.”
3. Recommendations for practice success – “Don’t rush to replace human agents with chatbots and other automation. Consider machine learning only when you have sufficient and high-quality data for the learning. Take the users’ expectations and habits into account.”
This recognition of AI’s growing importance to ITSM is great. Still, it does little to help ITSM practitioners adopt AI-based capabilities such as:
· Virtual agents or Chatbots that answer end-user queries and guide them through resolutions 24/7.
· Automated service level agreement (SLA) management – monitoring and helping ensure compliance with service-level targets, predicting potential SLA breaches, and automating escalation for high-priority tickets.
· Multilingual support using AI-powered language translation. For example, translating end-user queries and support responses in real-time.
2. Incident management
The latest incident management practice guide (accessed January 2025) does not mention artificial intelligence or AI. It does, however, have three machine learning references as follows:
1. General information section – “Automated incident detection and categorization may benefit from machine learning solutions, using the data available from past incidents, events, known errors, and other sources.”
2. General information section– “… possibility and quality of machine learning.”
3. Information and technology – “Leverage machine learning capabilities: Incident detection, matching, classification and prioritization can be enhanced or fully automated using machine learning. Effective use of machine learning requires high-quality data and effective integration with various sources of information. If used properly, it can significantly improve the incident management practice.”
As with the service desk practice guide, the incident management practice guide recognizes the importance of AI to ITSM, but it does little to help ITSM practitioners adopt AI-based capabilities such as:
· Handling end-user issues without the need for human intervention. For example, automatic user password resetting or troubleshooting end-user issues.
· Predictive analytics for incident prevention which involves anticipating and addressing issues before they impact end-users. For example, hardware failure can be predicted based on usage and performance data, or system logs can be monitored to detect anomalies that indicate potential issues.
· Enhancing the usability of knowledge bases with AI-driven search and recommendations. For example, suggesting relevant articles or solutions based on end-user queries.
3. Service request management
The latest service request management practice guide (accessed January 2025) does not mention artificial intelligence, AI, or machine learning. However, there are many possible AI use cases, including:
· The aforementioned virtual agent capabilities.
· Intelligent ticket routing, where tickets are automatically assigned to the most appropriate personnel. Workflow automation, such as automating software installations and updates or automating access management workflows, such as granting permissions.
4. Change enablement
The latest change enablement practice guide (accessed January 2025) does not mention artificial intelligence, AI, or machine learning. However, there are many possible AI use cases, including:
· Risk assessment and mitigation, evaluating the potential risks associated with proposed changes and predicting the likelihood of change failure.
· Change impact analysis capabilities to understand the potential effects of changes on systems, users, and processes.
Automated change prioritization based on business impact and urgency and intelligent change scheduling.
5. Problem management
There is one mention of AI in the problem management practice guide (accessed January 2025) as follows:
1. Recommendations for practice success – “Use AI and automation tools where possible. AI can interrogate data to identify trends. For example, AI and RPA can be used to remove repeat and underlying issues.”
There are also two references to machine learning:
2. Information and technology – “Machine-learning-based problem identification based on analysis of past and ongoing incidents. Management of problem records integrated with other service management data.”
3. Information and technology – “Leverage machine learning capabilities: Problem identification based on incident records or on monitoring data; cause-effect analysis; impact analysis, search for resolutions, and other activities of the problem management practice can be automated with the use of machine learning. Effective use of machine learning requires high-quality data and effective integration with various sources of information. If used properly, it can significantly improve the problem management practice.”
Again, the problem management practice guide recognizes the importance of AI to ITSM. Still, it does little to help ITSM practitioners adopt AI-based capabilities such as:
· Root-cause analysis (RCA) automation to identify the underlying cause of recurring problems.
· Problem pattern detection such as grouping similar incidents using AI clustering algorithms.
Predictive problem management to help anticipate and mitigate potential issues before they occur.
ITIL 4 and AI
As the five example ITIL 4 practice guides show, AI was still far from gaining mainstream traction back in 2019 (when the practice guides were written). However, while recognizing the importance of AI to ITSM capabilities, there’s nothing to explain how such capabilities should be adopted within these practice guides.
It could be argued this makes sense – that it would be better to include this guidance in a separate publication that references use cases with in these five and other ITIL 4 practice guides. There was a 2019 “ITIL 4 and Artificial Intelligence” paper available in the old Axelos subscriber portal, but it is no longer accessible in the PeopleCert subscriber portal.
ITIL 4 knowledge management global best practice
The ITIL 4 knowledge management practice guide wasn’t included in the above service desk management practices, but it’s interesting to see not only more mentions but also the call out for generative AI in its text:
1. General information – “Employees should be taught to discover and process information in the most efficient and valuable way. This includes training in using emerging information sources, such as large language models and other generative AI tools.”
2. Information and technology – “Embrace AI: utilize generative AI, large language models, and other AI technologies, alongside data science, to enhance various aspects of knowledge management. Leverage AI to improve access to knowledge bases, refine methods of information processing and analysis, and facilitate more effective knowledge sharing and generation.”
It’s likely a sign that the knowledge management practice guide has been updated more recently than the others. But again, there’s no guidance on the what and how for IT practitioners to leverage.
What this means
If your IT organization is looking to leverage AI to improve its IT service delivery and support, the ITIL 4 practice guides, while offering valuable guidance for specific ITSM capabilities, do not offer actionable specifics on AI adoption (yet, and expect to see some related guidance in 2025).
Instead, the best guidance for AI adoption in ITSM use cases comes from organizations that have already successfully embedded it into IT operations. This is either peer organizations or the suppliers of AI-powered ITSM tools, with the latter more likely to be better prepared to advise your organization on the what and how of AI adoption based on multiple successes across a spectrum of organizations.