What Now Assist is
Now Assist is a set of AI capabilities embedded directly in the ServiceNow platform — not a separate product you install. It uses generative AI (built on LLMs including models from Microsoft Azure OpenAI) to generate text, summarise content, suggest next steps, and write code.
Now Assist for ITSM
Incident Summarisation: Reads the work notes and activity stream on an incident and generates a plain-language summary of what has been done and the current status. Saves L2/L3 engineers from reading 50 work notes to get up to speed.
Resolution Notes Generation: Suggests a resolution note based on the incident details and work notes. Agent clicks accept or edits before saving.
Chat Assist: Virtual agent conversations powered by AI — can resolve common requests without human intervention.
Now Assist for Developers
Code Generation: In the Script Editor, type a comment describing what you want and Now Assist suggests the code. Works for GlideRecord queries, Business Rule logic, Client Scripts.
Code Explanation: Select existing code and ask Now Assist to explain it in plain English. Useful for understanding inherited code.
Test Case Generation: Generates ATF test cases based on your script logic.
AI Search
Replaces keyword-based search with natural language understanding. Users type a question and get relevant knowledge articles rather than a list of keyword matches.
Licensing reality
Basic Now Assist capabilities come with the base platform (Washington DC and later). Advanced capabilities — including higher usage limits, custom LLM configurations, and some HRSD features — require the Now Assist add-on licence. Check your entitlements before promising specific features to stakeholders.
Limitations to know
- Generated code must be reviewed — Now Assist does not know your specific instance's custom tables, fields, or business logic
- Summarisation quality depends on work note quality — garbage in, garbage out
- Not all languages are supported equally in AI Search
- AI-generated content requires human review before use in production automation