Decision automations and decision services
Decision automation
A decision automation is a container for Decision Designer users to organize one or more related decision services. When you open a decision automation, Decision Designer automatically loads all the decision services it contains.
Each decision automation must be connected to a remote Git repository (remote repository) to share the decision automation collaboratively. When you create a decision automation, you associate it with a branch that is synchronized with the repository. A branch contains a version of all the files and their change history. You can create several branches.
To understand branches, take as an example the operations of a bank. A bank might reassess offerings such as loans and corporate policies several times a year. You might want to create a branch each time a new policy or a new loan rate goes into effect.
Decision service
After you create your decision automation, you can create decision services in Decision Designer. A decision service contains decision artifacts that implement a decision:
- Decision model
- The representation of a decision. It contains a diagram that describes a decision, its subdecisions, the data that is required to make these decisions, and how they are related to each other. It includes rules and decision tables that describe the decision logic.
- Data model
- The representation of business concepts in the form of data types. Use data models to describe all the data that you need to make your decisions.
- Predictive model
- The representation of a prediction based on a machine learning model. Predictions are consumed by decision models and task models.
- Task model
- The implementation of a decision. It contains rule artifacts and a ruleflow to organize their execution.
Continuing the bank example, you might have a decision service for mortgages, and another one for student loans. The decision service for mortgages might contain a decision model for fixed rate mortgages and another one for adjustable rate mortgages. This decision service might also include a data model that represents, for example, the customers and their personal data, financial documents, and available mortgage options.