Creating, deploying, and monitoring a decision service

This tutorial guides you through the process of creating, sharing, deploying, and monitoring a decision service in Decision Intelligence.

Learning objectives

Decision Intelligence includes Decision Designer, a web-based development environment for creating decision services. In this tutorial, you will learn how to:

  • Create a decision automation and a decision service.
  • Build a decision model and define its logic.
  • Run the decision model with test data.
  • Share your decision service.
  • Select collaborators for the decision automation.
  • Create a data model and link it to the decision model.
  • Create an operation and version your automation.
  • Deploy the decision service and run it in the decision runtime.
  • Create a dashboard and charts to monitor the execution of your decision service.

Key concepts

You start by creating a decision automation that contains a decision service. A decision service uses decision artifacts to define business decisions:

  • Decision model: A diagram with nodes for decisions and data inputs. Decision nodes contain logic that processes input data. The diagram controls data flow among nodes. Predictive and function nodes can call other models defined in the service.
  • Task model: Chains tasks together and defines conditions for execution.
  • Predictive model: Applies data from a machine learning model.
  • Generative AI model: Analyzes or generates content based on context.
  • Data model: Defines custom data types used in the decision service.

This tutorial focuses on creating and using a decision service based on a decision model. In the final task, you also learn some of the concepts needed to monitor your deployed decision service by using dashboards:

  • Monitoring source: Identifies the deployed decision service whose execution data is shown in dashboard charts.
  • Aggregation: Defines how values are calculated in charts, such as Count to total the number of executions.
  • Group by: Splits results by a field, such as grouping executions by stormAlert.
  • Visualization type: The chart format used to display results, such as gauge, pie, and so on.
  • Filter: A condition applied to the dashboard that limits the data shown in charts to items that match a selected field and value.
  • Time range: Defines the window of time used to include or exclude execution events in charts.
  • Thresholds and alerts: Options in Period KPI charts that let you set a limit value and trigger an alert when it is reached.

A Getting started sample is available under Discovery tutorials in the New decision service wizard. The decision service has a separate decision model for each task in the tutorial. You can use the decision service to tour the product, or run it alongside the tutorial to work through the tasks. For more information about importing the sample, see Building decision services.

Audience

This sample is designed for anyone who wants to learn how to use Decision Intelligence.

Time required

35 minutes

Prerequisites

You need access to Decision Designer.

Tasks