Mapping Participants' or Collateral's evaluation algorithm
Last updated
Was this helpful?
Last updated
Was this helpful?
Mapping Decision Algorithms to Participants and Processes
Decision algorithms can be customized within the framework of defined processes and Participants or Assets (like Collateral). For example, in Scenario 3.2, the scoring process is split into two stages: an initial "SoftHit" and a final "HardHit."
In the SoftHit phase, the decision algorithm operates solely on participant information without using external data.
The HardHit phase is triggered when the participant consents to external data queries, allowing additional business rules to be included in the evaluation.
Assigned BPMs:
Flowable Requests and Customization of Scoring Stages:
If you don’t see the required process or participant on the mapping page, you may need to request Flowable for the specific status. For this, refer to:
Flowable (BPM) method request
The number of potential scoring stages is unlimited and can be aligned with the bank’s processes. Before deploying a new decision algorithm, users can test it against pre-existing data in the system.
Attributes generated by the decision algorithm help build the client profile, influencing the offers available to the client. These attributes may include CreditScore, Net Disposable Income, Fraud Index, and Gross Income.
However, not all client profile attributes come from the decision algorithm. Depending on the business logic of the process, some attributes might result from user actions. For example, if a borrower disputes the offered amounts and provides details of additional income sources, this information, once verified by a designated user, becomes part of the client profile, prompting the system to automatically update the offers.
The execution of an algorithm results in:
The creation of a defined client profile for a specific participant.
A set of warnings to categorize the client within the yellow zone.
A list of stop-factors for red zone placement, leading to an application decline.
These warnings and stop factors are displayed within the system, providing context.
The outcomes from the exectution are also saved in the Client Profile section, and later can be used as variables for pricing tool, product segmentation, documents and more:
The profile parameters can be very sophisticated, as seen in the SMB case:
Additionally, the system visualizes data obtained from each step of the algorithm's execution, showing the details for every data source involved. An executed decision algorithm for a participant may trigger a warning or alerts, like this (learn more about it and ):
