Implementation of the Credit Decisioning System based on Alphyn ASM at AVO Bank
- Customer
- AVO bank
- Project manager on the customer side
-
- IT Provider
- Databorn
- Year of project completion
- 2025
- Project timeline
- February, 2025 - November, 2025
- Project scope
- 2400 man-hours
- Goals
-
The goal of this project was to implement a real-time decision-making system for credit issuance at AVO Bank. The bank had only recently entered the market, and already had a system that calculated personalized terms for credit cards but it had been built in-house, under tight time constraints.
During the project, the old system was replaced with a new one based on the Alphyn ASM (Analytical Stream Manager). The existing checks and calculations were migrated from the old system to the new one, and updates were made that had not been possible to implement in the old system. - Project Results
-
The new decisioning system has successfully replaced an old one. The bank continues to grow and evolve, and in the near future. In other words, in less than a year, the new system based on Alphyn ASM has already surpassed the old one in terms of functionality.
The uniqueness of the project
The Alphyn ASM was originally designed with maximum flexibility in mind. On the one hand, thanks to its user-friendly graphical interface, it allows business users who aren’t familiar with programming to make changes to the business process. On the other hand, the platform’s support for Python, as well as Java and Groovy, enables experienced developers to implement within the system any business logic — including highly complex scenarios — required by the bank.
Bank’s credit managers are very proactive, so the credit policy is frequently updated: the checks / rules that loan applications go through are regularly changed, new ones are created, and old ones are deactivated or updated. The business might unexpectedly demand that groups of checks and calculations be rearranged arbitrarily. During the project, the new system was configured from the ground-up with the expectation of frequent future changes to the business processes.
- Used software
-
Alphyn ASM is a Docker Container-based solution that was built using the following software:
● Backend: Python, Java, Flink● Frontend: Nginx, React● Database: PostgreSQL● Integration: Apache Camel, Kafka
- Difficulty of implementation
-
The most challenging part of the project was developing integration mechanisms. During development, it became apparent that the external data sources (for example, credit bureaus) supplying the system with necessary information are often unstable. A response to a request might arrive in a few seconds, a few minutes — or a few hours. A standard decision-making approach would not be suitable in such a case: the number of unjustified denials due to missing data would be unacceptable.
With this in mind, the analysts designed a specific process flow using Alphyn ASM built-in capabilities. The new flow allows each credit application to wait for the required data and receive a properly determined decision — no matter how much time it takes.
- Project Description
-
Alphyn ASM provides a graphical interface to build automated decision-making strategies for various types of loans. A process flow for credit cards was developed as the first strategy during the project.
Loan requests are collected through multiple channels, such as the bank's mobile app or the website.
Requests are processed automatically: all the checks / business rules and all the integrations are automated. The system gathers all the necessary data from the external services (such as the bank’s own databases, credit bureau, blacklists, etc.), performs the necessary checks and calculates the final decision.
If the customer is eligible for a loan, the final offer (rate, loan amount, loan term) is calculated and sent to the CRM and to the customer.
- Project geography
-
AVO Bank is headquartered in Tashkent, the capital city of Uzbekistan. The bank provides its services online (through their app and the website) as well as its extensive network of ATM machines that are located in all major urban centers across the country.
Databorn is headquartered in Dubai, UAE, and has multiple offices in other countries, including 3 offices in MENA countries and 2 resource centers in Europe (Serbia) and CIS (Kazakhstan).