There’s a lot of interest in the opportunity of artificial intelligence (AI) for IT service management (ITSM), especially with the media hype caused by ChatGPT and other generative AI tools. This is not only influencing ITSM tool product development, it’s also had an impact on “marketing machines” and the terms employed to help sell products and services.
Two such terms are AITSM and AIOps, and it’s essential to appreciate that they are two distinct “things” despite their usage. To help, this blog looks at what AITSM and AIOps are, starting with their definitions.
AI-focused ITSM definitions
According to Gartner Research, “AITSM is not an acronym; rather, it is an initialism. It is a concept that refers to the application of context, advice, actions and interfaces of AI, automation and big data on ITSM tools and optimised practices to improve the overall effectiveness, efficiency and error reduction for I&O staff.”
Whereas AIOps (Artificial Intelligence for IT Operations) is defined in the Gartner Online Glossary as: “AIOps combines big data and machine learning to automate IT operations processes, including event correlation, anomaly detection and causality determination.”
On the face of it, these definitions might seem similar (based on the words they use). However, there are important differences between AITSM and AIOps. These are outlined below.
Key AITSM capabilities
The key thing to say about AITSM is that it relates to ITSM tools and the addition of AI-enabled capabilities to them to improve IT service delivery and support operations and outcomes. So, for example, capabilities such as:
- Ticket triage (categorisation, prioritisation, and routing) using “intelligent” workflows. For example, the ITSM tool automatically categorises a ticket as an “application issue”, and routes it to the appropriate team, based on the ticket text.
- Providing initial diagnostics and suggesting potential solutions or workarounds. This can include incident ticket matching.
- “Conversational” self-service using chatbots (or virtual assistants/agents) to retrieve relevant articles from the knowledge base(s), offer guided assistance, or access automation capabilities. For example, a chatbot interacts with the end-user to guide them through the steps required to reset their password.
- Automated email responses that share the most likely answers or resolutions to email enquiries.
- Identifying recurring issues and root causes for problem management, plus predicting issues before they are seen or even occur. For example, the ITSM tool forecasts a potential system crash and suggests preventative measures such as resource scaling.
- Trend analysis and impact and risk assessment for proposed changes based on historical data.
- Mapping dependencies between different configuration items (CIs) and automating CI updates in the configuration management database (CMDB).
- Creating new or updated knowledge articles based on closed-ticket text, curating and maintaining the knowledge base relative to use and feedback, and providing real-time knowledge-based recommendations to end-users and IT staff.
- Monitoring service levels and automatically generating service level management alerts and reports.
- Analysing customer feedback for service quality insights. Or using real-time sentiment analysis to identify potential issues based on the end-user’s IT support interaction, with escalation used if needed.
- Suggesting areas for service improvement.
Key AIOps capabilities
AIOps capabilities can be delivered via ITSM tools or other tool sets that are closer aligned with IT operations management (ITOM). The common AIOps capabilities include:
- Anomaly detection – automatically detecting abnormal behaviour or outliers in system performance. For example, IT staff are alerted to a sudden spike in server CPU such that the potential impact can be investigated and addressed as needed.
- Event correlation and analysis – aggregating and correlating event alerts from various sources to identify the most significant issues and eliminate the event “noise”.
- Predictive analytics and actions – using historical real-time data to understand and pre-emptively address potential future issues. For example, data patterns show that an infrastructure component will fail or perform unexpectedly, and the issue can be remedied before it impacts business performance.
- Performance monitoring – such as system health, where the performance and availability of IT services and infrastructure are continuously monitored. Or end-user experience tracking and analysis.
- Capacity and cost optimisation – understanding how demand for IT services changes over time. For instance, understanding seasonal peaks and troughs (based on historical data and current trends) to scale or shrink the available infrastructure and its costs accordingly.
- Root-cause analysis – using event patterns and service topologies to identify the root cause(s) of service issues and problems. For example, recognising that end-users are experiencing slow load times on a web application, finding and rectifying the issue(s) (perhaps the code, database performance, or network latency), and making the necessary changes to prevent a recurrence.
- Automation and orchestration – which includes automating routine tasks and self-healing (or auto-remediation) where the technology knows an issue requires fixing and the necessary fix (which is applied automatically and either with or without human involvement). For example, service downtime is detected out-of-hours, predefined actions such as service restarts are automatically performed, and an incident is logged for later review.
While there are overlaps between AITSM and AIOps, most of the AITSM use cases relate to operational improvement using intelligent automation and data. While most of the AIOps use cases improve insight and infrastructure availability using data and intelligent automation.
AITSM or AIOps – which should your organisaiton start first?
The short answer is yet another question, “Which will deliver the most benefits for your organisation?” For many IT organisations, this will mean the adoption of AI-enabled capabilities in their ITSM tool to first improve their ITSM operations and outcomes (with this helped by not needing to acquire or access other IT management tools). It also better places them to benefit from new AIOps capabilities. Whereas if AIOps is introduced first, while the greater insight and automation will help IT organisations, the full benefits might be diminished if operational ITSM inefficiencies are still present.
If you want to learn more about AITSM and AIOps, please contact our team.