1 HIGHLY INTEGRATEABLE SOLUTION FOR ANY DECISIONS
The strength of the application lies in its openness and flexibility. The Solution is capable of taking the input data necessary for the decision process from any existing systems of the bank and also from external data sources, while the final output generated as a result of the rating or scoring workflow can be used by any connected application.
The data from the source solution or individual parameter tables are evaluated, scored, aggregated and assessed by ApPello Decision Engine. XMLs arriving from the source solution are managed by the Engine and the processed data are sent back by the same manner. Having regard to the frequent regulatory and environmental changes, in case of the expansion of the XML the system is able to handle the assigning new tags to the new variables with the help of configurable interface.
As a result of the evaluation of the application, the Engine is able to provide clear answers and information of further workflow steps thus reducing operational risk and improving efficiency.
2 ONE SOLUTION FOR ALL DECISIONS AND ALL SEGMENTS
ApPello Decision Engine is a solution for all your decision points across your organization.
It is highly integrateable with loan origination, loan management, monitoring, debt collection and other solutions.
It is capable to serve both retail and non-retail segments. The risk conditions can be separated from each other, be associated with the given segment and work independently.
3 FOR ALL COMPLEXITY LEVEL - FOR A BETTER PORTFOLIO
The Engine supports complex calculations for non-retail and retail segment using the advanced statistical and mathematical functions of the system. Logit transformation or PD calibration can be quickly implemented by the risk experts of the bank. The system also manages the different steps of the PD rating calculation (model PD, overruled PD, decided PD, etc.) and expected loss calculations.
4 Flexible scorecard definition
The Decision Engine is responsible for scorecards and for their maintenance. Questionnaires and scorecards can be flexibly and periodically adjusted to reflect regulatory requirements or credit policy changes. Within the process, the users are able to parameterize all rules in parallel.
Scorecards can be defined for different client segments, for different loan products and also for different purpose for example pre-scoring, client scoring. Already existing scorecards can be copied and modified into different versions.
5 VISUAL EDITOR FOR CHANGES WITHOUT IT
The Decision Engine contains parameterized data models, processes and rules so for a new product or a client sub-segment further questionnaire and scoring models can be created by the Bank’s experts. When it comes to modification of a logic or parameter board, it is possible to add new parameters, clone, inactivate/activate or archive the old one without vendor or IT intervention thanks to the easy-to-use system.
It has never been so easy to orchestrate, test, evaluate the scorecard’s rules with the help of decision tree that visualizes the whole decision process.
The system’s calculation/evaluation logic is highly formable and able to manage several logics at the same time based on the versioning function. For example, in a campaign for the same product different interest rates or conditions are applied.
6 CHALLENGE YOUR CALCULATION MODELS
ApPello Decision Engine is a powerful tool to challenge your practices and build up different models that can be versioned and test-live models can be compared.
Champion / Challenger approach is supported. If there is a need for quickly deploy a business logic modification without knowing the impacts of the modified business logic vs the existing business logic it is possible to define a percentage by the user for the challenger logic. After this the system assigns each request (application / customers) to either the Champion (existing business logic) or the Challenger (modified business logic) on a random basis according to the defined percentage.
Testing module supports what-if analysis also with comparison of the defined what-if simulations.