The decision engine is a basic tool supporting credit and operational risk management in financial institution and shortening time of making various types of decisions. The engine supports processes related to analysis of economic and financial situation of the customer, its reliability, credit risk assessment, detection of fraud and abuse and automation of decision making. In this way, it reduces potential losses and increases effectiveness of the organization

  • How are such functions performed by a modern decision engine?
  • In what other areas does it find application?
  • Does it benefit from using the engine?
  • What are the costs of the Ferryt Decision Engine compared to other tools in its class?

The answers to these and other questions can be found in this article.


Advanced decision engines not only need to provide the ability to build decision models, but also process creation functions. This function provides organization and standardization of the entire decision logic process.

Along with this, the engine is expected to be easily integrated with various data sources and to have adequate performance and scalability. High performance of the decision engine should be a native feature of the engine and not just achieved by providing a powerful IT environment.

Another important feature is the use of the ‘low-code’ technique, a visual approach to creating decision flow in the engine. This involves implementing logical processes using diagrams. This allows business users without programming skills to build and modify solutions.

Decision Flow in Ferryt Decision Engine

For performing more advanced calculations and building complex algorithms, an intuitive formula language should be available, similar in simplicity to those used in the well-known Excel tool.

Meeting the requirements outlined above makes up the most important feature of a modern decision engine – its versatility. The engine must provide opportunities to be used in different areas and perform multiple functions.

The Ferryt Decision Engine, from DomData, has these characteristics. This solution has been successfully used for many years by, among others, the largest Polish banks. Let’s take a look at the possibilities offered by this decision engine.


The Ferryt Decision Engine can implement any scoring, rating, behavioral or fraud
and abuse detection models. The engine allows to build an appropriate workflow which not only calculates business rules but also automates decision making. In many implemented systems credit decisions can be made completely automatically.

A user-friendly and ergonomic engine environment, provides support for the full solution build cycle. The user can verify the performance of models at every stage of their lifecycle. With the built-in debbuger function, it is possible to easily analyze the flows and code of the created models.

The engine allows testing a priori. This involves running any number of test use cases
on test and/or real-world data before production deployment. For models that are already running, the engine allows you to run tests a posteriori based on historical data analysis.

Solution logic diagram

All elements of the solution are stored in a repository, which enables versioning, comparison of differences between versions and models, and individual control of access rights.


Parameters for the model are always up-to-date due to the fact that the engine retrieves them at the time of starting the conversion. This guarantees on the one hand a current and actual result and on the other hand enables easy recalculations. These functions are used in recalculation of pre-approved offers, calculation of customer limits or in generation of warning signals on specific data sets.

The flexibility of Ferryt Decision Engine-based solutions is provided by the ability to use different formats and data sources. When creating a model using webservice, the user can connect to it own databases or external data sources. The defined decision models are available in the form of webservices, the publication of which does not require additional work from IT departments and model developers.


The engine provides secure management of access permissions to model definitions and calls. Versioning and auditability of changes made ensure confidentiality of models used.

The use of the reporting module allows the tool to be used not only in the context of defining and running models, but also supports the process of monitoring them.

This makes the decision engine a part of the architecture in financial sector institutions, covering the tool requirements specified in the supervisory recommendations of the Financial Supervision Commission for risk control.


Application of a modern decision engine is not limited to support of credit and operational risk management. It significantly speeds up the process of analysis and consideration of credit applications, ensures consistency and repeatability of decisions and makes it independent of human errors.

Ferryt Decision Engine can also be used to manage parameters of credit products, to support debt collection, to create and manage pre-approved offers, to prevent loss of customers. A natural application is to automate the handling of the complaint process or support the robotization of back-office processes in an institution. It can also implement various other econometric models, as well as artificial intelligence models.

Formulas in Ferryt Decision Engine


Ferryt Decision Engine is not only a flexible, efficient and stable tool, the use of which does not require programming skills.
It is also a cost-effective solution from financial side. Attractive licensing system of the tool, fast implementation requiring low expenditures, low maintenance costs, prevention of financial losses translate into achieving high ROI from the engine implementation. The cost of using Ferryt Decision Engine most often compares favourably to other tools of this class.

All of this makes the solution created by DomData applicable to any financial institution and can impact the achievement of business objectives at an optimal risk level.

Read more about Ferryt Decision Engine