Decisimo decision engine allows for classic business rules implementation. The evaluations of conditions are number, text, and time-based.
The decision engine allows for multiple independent variables (sometimes called conditions) and multiple dependent variables (sometimes called actions). The conditions used for evaluation can be one of 11 ranging from classic equal or not equal, to closed/open intervals.
Data source connection
For enriching data used in decision flows, any type of REST API that responds in JSON or XML format, can be connected. In case that multiple calls are to be, engine allows to make the calls in parallel to reduce overall execution time.
The classic types of data sources used in decision flows are requests to credit bureaus, public blacklist, or third-party scoring/data providers.
If you have machine learning trained that you intend to use in decision flows, the engine allows for direct import of model. Simple dump of final model using pickle, importing the pickle file to the portal, and specifying libraries used, will allow you to call the model directly and make decisioning not only on classic rule-based logic, but also your bespoke model.
The strength in the Decisimo engine is also in the ability to build your decision flows using drag and drop workflow builder. All components such as models, rule sets, decision tables and data sources, can be used in multiple different decision flows.
The execution of parts of flows can be conditional based on values within flows using fork (branching). Forking is the way, you can implement champion/challenger, or skip execution of certain parts may you know the result is already final even without proceeding into other parts of the flow.
Decision intelligence portal allows to run a unit test - a test to evaluate whether rules, decision tables, or models are executing decisions as expected. This way you can use different types of inputs and see results.
We have a network of execution endpoints. Once you build decision flow, test it out, you can deploy it to any endpoint in one of our locations. This allows for deploying decision flows to a specific region and ensure that data are not leaving a specific jurisdiction/region.
In your system, you integrate the specific endpoint once and then deploy new decision flows instantly.