Agentic workflows, agents, and skills

  • Release version: Zurich
  • Updated May 7, 2026
  • 2 minutes to read
    • Summarize
    Summarized using AI
    This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete.

    Summary of Agentic workflows, agents, and skills

    Build Agent enables ServiceNow customers to create agentic workflows, agents, and skills tailored to their custom applications without starting from scratch. This capability transforms business requirements into configured AI artifacts by leveraging the existing app context such as tables, roles, business rules, and metadata. To use these features, customers must install Now Assist for App Engine and have access to ServiceNow Studio or the ServiceNow IDE. Access is also controlled by entitlements.

    Show full answer Show less

    Key Features

    • Skill Creation: Supports creating various tools including FlowActions, Now Assist Skills (inline script or explicit), SubFlows, and WebSearch.
    • Agent Creation: Enables adding catalog items, conversational topics, flow actions, Now Assist skills, record operations, scripts, search retrievals, subflows, and web search capabilities to AI agents.
    • Scope and Reuse: Only FlowActions and SubFlows within the app scope can be used; Build Agent reuses existing ones rather than creating duplicates.
    • Search Retrieval Tool: For agents requiring knowledge retrieval, Build Agent configures a search retrieval tool that queries existing search profiles and semantic search indexes (requires Zurich Patch 10 or higher).
    • Four-Phase Artifact Generation: The process includes Assessing data model readiness, Analyzing use cases, Designing architecture with quick-win recommendations, and Building agents and skills scoped to data model, roles, and ACLs.
    • In-App Agents: Work with any custom scoped application, producing more targeted artifacts when real data and roles exist. Without data or configuration, agents are less specifically scoped.
    • Governance and Deployment: Generated agents are automatically registered in AI Control Tower for centralized governance and deploy through standard update sets.

    Practical Benefits

    • Accelerates AI-driven automation by generating tailored workflows and agents aligned with existing application structures and security.
    • Ensures consistency and governance through automatic registration in AI Control Tower and standard deployment processes.
    • Supports a wide range of tools and agent capabilities, enabling versatile AI integrations within custom apps.
    • Improves relevance and effectiveness of AI agents by leveraging real application data, roles, and automation.

    Build Agent can generate agentic workflows, agents, and skills scoped to your custom app. Turn business requirements into configured AI artifacts without building from scratch.

    You must install Now Assist for App Engine to create agentic workflows, agents, and skills. For more information, see Installing Now Assist for App Engine.

    Verify that you have access to ServiceNow Studio or the ServiceNow IDE.
    Note:
    Your entitlements determine whether you have access to agentic workflows, agents, and skills in Build Agent.

    Skills must be published before they can be used as a tool by another skill.

    Build Agent supports the following tool creation:
    Note:
    1. *You can use only FlowActions and Subflows that are already part of your app scope. Build Agent reuses existing FlowActions and Subflows instead of creating new ones.
    2. **When you create an agent that requires knowledge retrieval, Build Agent can configure a search retrieval tool scoped to your use case. The tool queries your existing search profiles to identify the best match, retrieves the associated search sources and return fields, and checks whether semantic search indexes are available.
      Note:
      You must be on Zurich Patch 10 or higher to use the search retrieval tool.

    The agentic capability turns business requirements into fully configured agents, skills, and agentic workflows for custom applications. It evaluates existing app context, such as tables, roles, business rules, and metadata, and applies a four-phase pattern to generate artifacts scoped to your application's data model, roles, and ACLs:

    Assess
    Evaluates data model readiness for AI.
    Analyze
    Identifies use cases from existing tables, roles, and flows.
    Design
    Proposes an architecture with a recommended quick-win.
    Build
    Generates agents and skills scoped to the application's data model, roles, and ACLs.

    In-App agents can work with any custom scoped application. Applications with real data, established roles, and defined automation produce more targeted, domain-specific artifacts. Applications with no data or configuration produce agents that aren't scoped to a specific data model.

    Note:
    Generated agents are automatically registered in AI Control Tower for centralized governance, and artifacts deploy through standard update sets.