Key Capabilities:
Unified interface to enable, disable, and monitor Now Assist skills across personas and channels. Provides visibility into feature usage, licensing, and available updates tailored to CSM use cases.
Allows configuration of in-line generative options directly within record fields or lists, such as summarizing notes, rephrasing, or generating replies on demand. Offers localized control of AI entry points for case forms and workspaces.
Out-of-the-box control panel to manage AI features, toggle skills, and contextually answer questions
Monitor usage and performance of generative AI features used in Now Assist for CSM or created through Now Assist Skill Kit
Streamlined UI to create, customize OOTB and publish generative AI skills like summarization and recommendations tailored for case types. Click here a YouTube playlist for a deep dive on NASK.
Policy enforcement layer to define boundaries for what AI can generate, grounded in CSM use cases like case types, roles, and customer data.
Enables admins to bring their own foundational model (BYOM) to power Now Assist features. Admins can configure external LLM endpoints (such as Azure OpenAI, AWS Bedrock, or private models) and manage routing, authentication, and security policies for CSM generative experiences.
CSA allows admins to define reusable, low-code subflows and associate them with natural language intents. These can be triggered from the Now Assist virtual agent or workspace panels to automate common agent actions (e.g., refund initiation, entitlement checks) without scripting. Especially valuable in CSM to accelerate resolution through AI-triggered automation.
-
Now Assist in Document Intelligence: Enables retrieval-augmented generation (RAG) by indexing CSM-specific documentation, forms, and attachments for use in real-time prompts. This is a generative AI version of its predictive AI predecessor.
-
Knowledge Graph Connector: Links your CSM environment to the Knowledge Graph, enabling semantic understanding of case data, taxonomy, and relevant content such as CRM Foundational data models.
-
GAF (Group Action Framework): powers intelligent case resolution by using LLMs and clustering similar records (Cases, Knowledge, etc) and applying generative AI to suggest relevant next steps or actions. It underpins features like Suggested Steps by analyzing common patterns in historical resolutions and surfacing them in real time.
- ServiceNow Lens: use generative AI to scan, extract, comprehend, and synthesize data to optimize your workflows.
Visual low and no code environment to design, test, and deploy agentic workflows. Empowers CSM teams to create decision-aware, multistep AI agents that interact with customers and platform data through cases, knowledge, and automation.
Provides a curated dataset optimized for Now Assist training and evaluation. Ensures CSM-specific fields, records, and taxonomies are properly indexed, enabling grounded and context-aware AI interactions.
Unified oversight and governance dashboard for enterprise AI across ServiceNow. Monitors LLM usage, policy enforcement, skill performance, and model reliability, giving CSM admins actionable insight into AI operations.
Enable Now Assist capabilities in the user's preferred language
Key Best Practices
Now Assist Skill Kit (NASK) FAQ
Now Assist Skill Kit – Tool and Deployment Options
Comprehensive Resource List for Now Assist Implementations
Measured Success
How do we usually measure success in this set of purpose-driven skills?
| Outcomes | Explanation (with applicable use case) | Success Metric |
|---|---|---|
|
Faster time to production |
Reduce time to deploy AI capabilities using packaged skills and simplified admin interfaces Use case: Panel, NASK |
Avg. configuration time (days) |
|
Flexible model control |
Allow customers to connect external LLMs while maintaining ServiceNow integration Use case: Generative AI Controller |
% of Now Assist experiences using custom LLM endpoints |
|
Scalable automation governance |
Centralize policy and signal management for AI decisioning across CSM workflows Use case: NAG, CSA |
% of users governed by skill- or signal-based policies |
|
More reusable AI-driven flows |
Increase reuse of subflows that trigger from natural language in Virtual Agent or Now Assist Panel Use case: CSA (Conversational Actions and Subflows) |
# of CSA flows used across agents or entry points |
|
Better clustering-driven outcomes |
Use past case clusters to provide tailored resolution steps at scale Use case: GAF (Group Action Framework) |
% of cases resolved using GAF-powered Suggested Steps |
|
Improved retrieval and accuracy |
Enhance generative responses with indexed customer documentation and attachments Use case: Doc Intel, KG Connector |
Knowledge response precision or accuracy rating |
