Using agentic AI in ServiceNow Otto for Customer Service Management (CSM)

The Customer Service Management (CSM) AI Agent Collection provides a set of prebuilt, fully configured AI agents and agentic workflows designed to address common Customer Service Management scenarios.

These model examples combine autonomous and supervised flows to perform multi-step actions using advanced reasoning, triggered by customer cases, conversations, or detected intents. Built on the ServiceNow AI Platform, the agents leverage capabilities such as Knowledge Graph, Flow Designer, scripting, Topics, Catalog Items, and Retrieval-Augmented Generation (RAG). They also use record operations, web search, and generative inputs. Powered by AI Agent Fabric and Workflow Data Fabric, the collection operates seamlessly across systems without dependency on data location. Guardian enforces guardrails for security and compliance. By offering ready-to-use building blocks, it helps guide and automate complex processes, reduce agent friction, and free human agents to focus on higher-value work. This accelerates resolution times and enhances customer experience.

Table 1. Available agentic workflows for Customer Service Management agent collection

Agentic workflow	Description	Available AI agents
Triage cases	Handles end-to-end case or case type validation, creation, verification, and escalation. It can also retrieve relevant context and details from the provided case or interaction to address customer inquiries directly and avoid unnecessary case creation.	Triage cases WrapUp Context validator Informational queries Case creation Entity extraction Document verification Email response
Provide customer 360 insights	Provides agents with real-time, context aware responses to queries on customer data, case details, product information, catalog entries, and interaction history, using multi-turn Q&A to maintain conversational context and accuracy.	Customer insight AI Agent Case action AI Agent
Complaint Case AI Agent collection	Automate and enhance the complaint resolution process by gathering missing information, detecting customer sentiment, categorizing complaints, and proposing resolutions. It supports human agents by managing complaint intake, triage, research, resolution, and ongoing communication, reducing manual effort and case closure time.	Complaint case intake agent Complaint case triage agent Complaint case research agent Complaint case summarization skill

Supported Large Language Models

Note:

You can use Now LLM Service, Azure OpenAI, Google Gemini or Anthropic Claude on AWS as the AI model provider for all generative AI skills and AI agents. Use the Configuration Controls in AI Control Tower to define which options are available, then set the skill-level preferences in the AI Admin Hub console. For more information, see Large language models on the ServiceNow AI Platform®.

Security implementation considerations

Enable security implementation to execute AI agents and agentic workflows through Access Control Lists (ACLs) and user identities. For more information, see Implement access control in AI agents

Considerations for running the autonomous AI Agents

Important:

By default, all agentic workflows and AI agent records are read only.

To run the AI agents autonomously, you must first duplicate the agentic workflow, and then proceed with the following steps:

• Activate the agentic workflow.
• Activate all agents within the agentic workflow.
• Activate the trigger to invoke the agentic workflow automatically. If you prefer to invoke it manually, activating the trigger isn’t necessary.

Standalone AI agents

There might be AI agents installed on your instance that are not used in agentic workflows. To learn how to see all agents that are available to you, see Find AI agents.