On 6 June 2025, ServiceNow released Now Assist for IT Service Management, its first attempt at embedding generative AI directly into standard ITSM workflows. Rather than a radical reinvention, this tool takes familiar modules—incidents, problems, changes—and layers on AI-driven summaries, suggestions and draft automations.
Core capabilities
Incident and Problem Summaries: Pulls together conversation highlights and relevant metrics into a brief, readable synopsis.
Resolution Recommendations: Offers potential fixes drawn from your own knowledge base and anonymised best-practice datasets.
Automated Drafts: Generates starter code or workflow templates for common repetitive tasks, such as password resets or routine patch scheduling.
These features are not intended to replace human judgement. Instead, they aim to reduce the manual overhead of information gathering and initial drafting.
The case for tempered expectations
Generative AI in ITSM often gets hyped for “fully autonomous” operations. In reality, organisations benefit most when AI handles the groundwork—summaries, first-pass drafts—while experienced practitioners validate and refine the output. Early adopters report modest time savings (around 20–30%) on standard ticket types, but the real gain is consistency in documentation and fewer basic follow-up queries.
Adoption considerations
Data Privacy and Compliance: Ensure all sensitive fields are masked or excluded before training and inference.
Governance Frameworks: Define clear approval processes for AI suggestions, with audit trails in place.
Pilot Scope: Start with low-risk service requests or static knowledge articles before expanding into change management.
User Sentiment: Provide training sessions so teams understand what the AI can (and can’t) do.
Early insights from trials
In a small‐scale trial at a regional university, agents using Now Assist saw a 25% reduction in time spent drafting incident descriptions, with minimal change to overall ticket quality. Feedback highlighted that summary accuracy varied by ticket complexity—simple password resets worked well, whereas multi-step network issues still needed substantial human input.
Looking ahead
As generative models become more reliable and organisations mature their AI governance, expect incremental improvements rather than overnight revolutions. Potential next steps include:
Proactive Alerts: AI-driven anomaly detection that suggests preventative actions.
Automated Change Assessments: Initial risk analyses drafted by AI for review by change advisory boards.
Cross-departmental Expansion: Similar AI assistants deployed in HR or facilities management, promoting a unified employee experience.
Conclusion
Now Assist represents a cautious, incremental move towards generative AI in ITSM—focusing on co-pilot functionality rather than full automation. For IT leaders, the real question isn’t whether to “do AI,” but how to integrate these tools into existing governance, measure their impact on agent productivity and ensure that human expertise remains central to every decision.