AI in ITSM has shifted from promise to practice. Twelve months ago we were playing and experimenting, and now for service teams trying to stay relevant, it’s no longer a question of if AI will play a role, it’s how and where. While the hype can be hard to cut through, there are real, working use cases in play today that are delivering measurable value. This post looks at where AI is genuinely helping IT Service Management and why it’s worth paying attention now.
Where AI Is Already Making an Impact
1. Incident Categorisation and Routing
AI models are now commonly used to categorise and triage incidents in real time. Instead of relying on manual assignment, machine learning uses patterns from historical tickets to route them to the right resolver group.
Current Leaders
ServiceNow
ServiceNow’s Predictive Intelligence uses supervised machine learning to analyse historical incident data and automatically recommend categories, priorities, and assignment groups. It improves over time as more data is collected.
BMC Helix ITSM
Offers cognitive automation with integrated BMC Helix Cognitive Service Management. It categorises incidents, routes them, and can even trigger fulfilment actions based on context.
Ivanti Neurons for ITSM
Ivanti is building momentum with its Neurons platform, focusing on hyper-automation and contextual routing. It learns from patterns in user behaviour and historical resolutions.
Why it matters
Faster response, fewer hand-offs, less frustration for both end users and support teams.
2. Predictive Outage and Alert Analysis
By analysing log data and telemetry from systems, AI tools can surface early indicators of potential outages or performance degradation, often before a ticket is raised.
Current Leaders
ServiceNow
ServiceNow’s AIOps (ITOM Predictive AIOps) module combines event management, telemetry ingestion, and machine learning to detect anomalies and correlate alerts to business services.
Dynatrace (with ITSM Integration)
Dynatrace is an observability-first platform with strong AIOps. While not an ITSM tool itself, it integrates seamlessly with tools like ServiceNow and Jira to raise predictive incidents.
Why it matters
Shifts the model from reactive firefighting to proactive issue prevention.
3. Knowledge Article Recommendations
AI is being used to suggest relevant knowledge base articles both to agents and end users. This can reduce time-to-resolution and boost self-service adoption.
Current Leaders
ServiceNow
Offers contextual knowledge suggestions through its Agent Assist and Virtual Agent features. When an incident is logged, the system uses historical resolution patterns and NLP to recommend relevant knowledge articles.
Freshservice (by Freshworks)
Uses Freddy AI to surface article suggestions as tickets are submitted or viewed, supporting both agents and requesters.
Atlassian Jira Service Management + Confluence
While not as advanced in AI as ServiceNow, Jira integrates well with Confluence to show article suggestions to users as they raise tickets or browse the help centre.
Why it matters
Accelerates resolution times and increases the value of existing documentation.
4. Automated Change Risk Assessment
AI is being applied to assess the potential risk of a change request based on type, past outcomes, impacted services, and even time of day.
Current Leaders
ServiceNow
Predictive Intelligence assesses risk scores using past change data and business impact analysis. The Change Success Score feature is especially useful in CAB meetings.
BMC Helix ITSM
Leverages machine learning to analyse historical change data and flag risky change types based on patterns of failure or delay.
4Me
Offers a more lightweight but still effective change risk predictor using change history and service dependency mapping.
Why it matters
Supports safer, faster decision-making in Change Advisory Boards and change workflows.
5. Sentiment Analysis on Tickets and Feedback
Natural Language Processing (NLP) models are now able to gauge user sentiment from ticket descriptions or surveys, helping Service Management teams identify when tone doesn’t match priority.
Current Leaders
ServiceNow
Provides sentiment scoring via NLP, allowing analysts to see emotional tone in ticket descriptions. This can be used to trigger escalation rules or flag for QA.
Zendesk
Uses AI to classify customer sentiment at scale and apply it to SLAs, routing, or support workflows, particularly strong in high-volume environments.
HaloITSM
Includes a feedback analysis module that categorises satisfaction scores and written comments for insight into agent and service performance.
Why it matters
Provides an early signal for service dissatisfaction and helps prioritise follow-up actions.
Not Every Use Case Is Ready and That’s Fine
It’s important to note that not all AI applications in ITSM are mature or appropriate yet. Chatbots, for example, are improving but still struggle with nuance and context in many environments. The goal is not to adopt everything at once but to identify where AI can improve clarity, speed, or decision-making and start there.
Final Thoughts
For many teams, the best AI use cases in ITSM are the ones that quietly remove friction, not the ones with the flashiest pitch. If it saves time, improves accuracy, or reduces rework, it’s worth exploring.
You don’t need to overhaul your entire toolset to get started. Most ITSM platforms now offer built-in or add-on AI features that can be piloted with minimal disruption.
FAQ Questions
What are the most useful AI use cases in ITSM today?
The most effective use cases include incident routing, predictive alerting, knowledge article suggestions, automated change risk scoring, and sentiment analysis. These areas offer measurable improvements to response time and decision making.
Which ITSM tools offer the best AI features right now?
ServiceNow leads in several categories including incident routing, knowledge suggestions, and AIOps. BMC Helix, Dynatrace (via integrations), Freshservice, and Zendesk also offer strong AI features for specific use cases.
Can AI improve change management in ITSM?
Yes, AI can assess change risk using historical data, type, timing, and affected services. This helps Change Advisory Boards make faster, safer decisions with fewer unplanned failures.
Is AI in ITSM only useful for large enterprises?
No. While larger platforms offer more advanced AI, smaller tools like Freshservice and HaloITSM are making these capabilities more accessible for mid-sized organisations.
🔗 Further Information
- ServiceNow Predictive Intelligence
- Dynatrace AI Powered Observability
- BMC Helix ITSM
- Ivanti Neurons for ITSM
- Dynatrace (with ITSM Integration)
- Freshservice
- Atlassian Jira Service Management + Confluence
- AI Capabilities and Adoption Rates in Leading ITSM Tools
- Predictive AIOps | The Next Layer in IT Service Management
Isnt all this AI stuff just about cutting jobs? It feels like we are automating roles people have spent years building skills and expertise in. Its sad.
I think thats a valid concern, and one many teams are either asking quietly or chatting amongst themselves, or maybe even directly. But in practice, and you need to ask, maybe AI in ITSM isn’t replacing people directly yet, it’s removing the repetitive tasks that burn them out. Categorising tickets, manually routing incidents, or trawling through knowledge articles, i know from experience aren’t the things most skilled IT professionals want to spend their day doing. The best AI implementations should free up service teams to do higher-value work like managing complex problems, refining services, or improving change processes. In fact, as these tools mature, demand often increases for people who can manage automation, interpret insights, and drive continual improvement.
So rather than eliminating roles, AI is changing them. It’s less about reduction and more about evolution. I would also add though that in my opinion, higher value work has become vague and overused. Is it strategic planning? Problem-solving? User engagement? Without specifics, it becomes a placeholder for something better without substance. Everyone from software vendors to consultants uses it to justify automation, cost-cutting, or restructuring, regardless of the outcome. Over time, it’s lost its meaning because it’s applied to nearly everything outside of repetitive tasks. In some of the places i have consulted at ive noticed in some circles, especially where I see job roles are changing, it’s perceived as a euphemism, a way to frame “we’re automating your work” in a more palatable way. That can cause skepticism. We should start saying, “let teams focus on solving problems, improving services, and making decisions that need human insight.” Lets be honest and think about it.
i understand the benefits but question how much trust we are placing into AI models that rely on historical data. If that data was biased, incomplete, or based on flawed process and lets face it data can be bad within IT, arent we just codifying bad habits? What happens when AI keeps recommending the same crap, incorrect categories or ignoring an edge case because it doesnt fit the pattern? Dont know about you, it just feels risky
I can only agree, and this is probably something that doesnt get tabled enough. AI models are only as good as the data and assumptions they are trained on. if there is bias or poor quality ticket history, it 100% can reinforce mistakes or blindspots. Ive seen this where AI just repeated outdated categorisations because that is what it learned. I guess if we treat the AI recommendations as just that, recommendations, then they should support the analyst, not replace them 🙂 and its essential to keep humans in the loop, especially when the AI is still learning. Some platforms like ServiceNow and BMC actually let teams review and correct AI decisions, which helps retrain the models and improve outcomes. So agree, its not set and forget at the moment, it needs governance, oversight and ongoing tuning just like any other part of ITSM. I dont think its about blind trust, its about helping humans make better decisions with a little less grunt work.