Add tools and information to an AI agent

  • Release version: Australia
  • Updated November 23, 2025
  • 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 Add tools and information to an AI agent

    This guide explains how to enhance AI agents in ServiceNow by adding tools and information that enable them to perform specific tasks effectively. Tools empower AI agents with necessary capabilities, while knowledge graphs provide contextual understanding of real-world relationships, improving AI outputs.

    Show full answer Show less

    Key Features

    • Tools Integration: Tools should be designed to complement each other and enable the AI agent to accomplish a defined objective. Detailed tool descriptions are crucial as they guide the AI agent in understanding each tool’s function and output handling.
    • Output Handling: When creating tools, consider how outputs will be processed, especially for tools that return large data sets. Use description or output transformation fields to specify processing strategies.
    • Knowledge Graphs: Adding knowledge graphs helps AI agents comprehend relationships between entities, such as mapping users to locations and departments to tailor processes like approvals.
    • Supervised Execution Mode: For sensitive or critical actions, tools can be configured to run under human oversight (supervised mode). This improves security and reduces risk by requiring human approval before executing certain actions.
    • Compliance Considerations: When using external tools like Google web search, be aware of data processing locations and organizational data policies, since Google routes requests through global datacenters, which may differ from your ServiceNow instance’s data region.

    Practical Application

    ServiceNow customers can use this process to tailor AI agents by selecting and configuring appropriate tools and knowledge graphs that precisely fit their operational needs. Ensuring tools have clear descriptions and compatible outputs improves agent performance and reliability.

    Enabling supervised execution for tools handling critical tasks adds a layer of security by involving human oversight, which is essential for sensitive operations.

    After configuring tools and knowledge graphs, customers can save their setup and proceed with guided AI agent configuration to deploy robust, goal-oriented AI agents within their environment.

    Add a tool to an AI agent to enable different functionalities and help your AI agents achieve their objectives.

    Tool overview

    Tools provide your AI agents with the capabilities necessary to complete their tasks. When adding tools, consider how the AI agent uses them to achieve its objectives and how those tools interact with each other. Providing your AI agents with the appropriate tools help promote the robustness and quality of their performance.

    Some additional guidelines for adding and creating tools are the following:

    • Design tools to work together. Your AI agent should solve a specific, discrete task, and the tools should give the AI agent the capabilities necessary to achieve its goal.
    • Write detailed tool descriptions. Tool descriptions are used by the AI agent to determine a tool's function. Thorough descriptions about what the tools are and how they work give the AI agent the best chance to succeed.
    • Consider the outputs of tools when creating them. Use the tool description or output transformation strategy fields to describe how to process the tool outputs. For example, if you have a tool that gathers records from many tables, provide the tool with plans for how to handle the large number of records.
    Note:
    If you select Google as your web search tool provider, the web search tool leverages Grounding with Google Search, offered under a Global Standard deployment. Because grounding isn’t data resident, Google's global infrastructure routes traffic to a global datacenter for each web search request. This processing may be different than your data processing location chosen for your ServiceNow instance. Consider your organization's data policies before enabling AI agents that use Google web search tools.

    Once you have added the tools to your AI agent, you can select Save and continue to move to the next step in the guided setup.

    Knowledge graphs

    In this step, you can also add Knowledge graphs. Knowledge graphs give the AI agent information to understand the relationships between real-world entities to improve its outputs. For example, you could add a Knowledge Graph to an approval AI agent that maps users to their location, company, and department to help the AI agent understand the specific approval process to apply.

    Knowledge Graph options

    Supervised execution mode for AI agents

    You can minimize the potential negative impact of an AI agent not executing as expected by configuring AI agents' tools to run in supervised mode. Running in supervised mode means that the tools use human oversight when executing actions. You can use the Supervised mode to enhance security for agents with the capability to perform sensitive or critical actions.

    You can set the supervised execution mode when creating a tool in the AI agent guided setup. For example, choose Supervised as the Execution mode when adding a catalog item tool. For reference, see Add a catalog item to an AI agent.