Decision tree nodes

  • Release version: Yokohama
  • Updated January 30, 2025
  • 3 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 Decision tree nodes

    Decision trees in ServiceNow consist of various node types—start nodes, question nodes, guidance nodes, and linking nodes—that guide agents through problem-solving workflows. These nodes work together to present questions, provide recommendations, and connect multiple decision trees for complex scenarios.

    Show full answer Show less

    Start nodes

    The start node initiates a decision tree by presenting the first question(s). When creating a decision tree, you define names, a title, description, and an action label that generate the start node. It always contains at least one question and a task input referencing the record type the agent is working on, enabling reuse in other nodes.

    Question nodes

    Question nodes ask one or more questions to evaluate a customer's situation. Answers determine the next node, which can be another question node, a guidance node, or a linking node. These nodes support HTML formatting for content enrichment. Using supported answer types is crucial to avoid runtime errors.

    Guidance nodes

    Guidance nodes offer recommendations or solutions, such as referencing knowledge articles, creating work orders, or escalating cases. They can only be followed by question nodes. Guidance nodes can define output variables that influence subsequent decisions and map outputs to the next node’s inputs. Authors can create custom guidances to reuse across decision trees.

    Linking nodes

    Linking nodes reference and connect to other decision trees, allowing workflows to span multiple decision trees, which is especially useful for complex or multi-stage troubleshooting. They include input mappings from prior nodes to the linked tree’s start node, maintaining continuity in execution.

    Agent navigation

    Agents can navigate backward to previous nodes using the Back button or jump to any prior node via the Edit icon in the “View my responses” tab, facilitating flexible response modification during the decision-making process.

    Decision trees include several types of nodes, including start nodes, question nodes, guidance nodes, and linking nodes.

    A decision tree begins with a start node, which presents the first set of questions to an agent. The start node is typically followed by a question node, which asks another set of questions. A question node can be followed by another question node, a guidance node, which provides a recommendation to an agent, or by a linking node, which includes a reference to another decision tree.

    You can enable agents to modify their responses in previous nodes:

    • Agents can go back to the previous node by selecting the Back button.
    • Agents can jump to any previous node by selecting the Edit icon Edit icon) in the View my responses tab.

    For more information, see Create a decision tree in Core UI and Add questions or instructions to a decision tree.

    Start nodes

    A start node is the first node in a decision tree. When you create a decision tree, you provide internal and external names, a title and short description, and a name for the action label. After saving the record, the system creates the decision tree and the start node. You can use this start node to build the rest of the decision tree.

    A start node includes a minimum of one question. A start node always has a task input of reference type so you can reuse the task input in other decision or guidance nodes via linking inputs. The task input includes a reference to the table that stores the record type that the agent is working on.

    Decision or question nodes

    A question node asks one or more questions. The answer to the question leads to an outcome, which can be another question, a guidance for the user to follow, or another decision tree.

    A question node presents the agent with a question or instructions to follow to evaluate the customer's situation and provide the right solution. You can add HTML formatting, such as images, links, or text formatting to the question node. Examples of typical questions or instructions include:
    • For resetting a password: "What is the customer's account ID?"
    • For troubleshooting a faulty device: "What is the device type or model number?"

    The answers to the questions are used to create path conditions.

    For more information about the available answer types, see Answer types for questions.

    Note:
    Using unsupported input types can cause errors in the runtime experience.

    Guidance nodes

    A guidance node proposes a solution to a customer’s issue through a reference to a guidance, which provides a recommendation to an agent. Examples of a guidance include recommending a knowledge article, creating a work order, or escalating a case.

    A guidance node can be followed only by a question node. The execution of the decision tree continues after the agent performs the guidance, and the guidance history is marked as complete. You can use a question node to connect two guidance nodes.

    A guidance can have outputs, which drive further decision making. With a guidance, you can do the following:
    • Define output variables.
    • Define values for the output variables, which are mapped in the Automation Plan flow for the guidance by the Assign Guidance Outputs flow action.
    • Connect the guidance to the guidance node's next node.
    • Map the outputs of the guidance to inputs for the guidance node's next node.

    For descriptions of the available guidances, see Guidances included with Guided Decisions Experience application.

    Users with the Decision tree author role can also create guidances to reuse in decision trees or recommended actions. For more information, see Create a guidance in the Core UI.

    Linking nodes

    A linking node includes a reference to another decision tree. This type of node connects the execution of the primary decision tree to another decision tree when a defined condition is met.

    Linking to another decision tree is useful for complex troubleshooting or for troubleshooting that requires multiple stages.

    A linking node includes the following elements:
    • A linked decision tree, where the execution of the primary decision tree continues after this node is reached.
    • Input mapping, which maps the inputs from the nodes preceding the linking node to the start node of the linked decision tree.
    • A path, which connects the previous question node to the start node of the child tree.
    For more information, see the following topics: