A Custom ML-based Credit Union CRM Platform

Comprehensive and accurate client risk assessment is a foundation of business success for any financial organization, especially for loan associations, where the danger of fraud is maximized. And while this procedure is absolutely vital, if performed manually, it can be time-consuming and inaccurate, which may lead to poor decisions, risky transactions, and financial losses.To prevent that from happening, our client decided to invest in a multifunctional tool for loan associations that would streamline the workflow and facilitate data-driven decision-making. According to our client’s vision, the project should have included a CRM for credit union, a user-friendly online loan calculator,  a reliable customer evaluation tool that would help to automate the screening process, and a task planner for the loan associations’ employees.Since the beginning, our client has seen this project as intended for internal use only. Therefore, the visual component was not a priority. At the same time, the system had to be user-friendly and intuitive to minimize the time required for onboarding and introducing new employees to the platform.

While working on this project wasn’t particularly challenging for Light IT Global as our team has already had previous experience developing custom CRM for credit union, we had to consider multiple unique factors that apply specifically to this project. The software was developed from scratch, and due to the fact that it should have included Machine Learning elements, we utilized Python for back-end because it offers simplicity, consistency, easy optimization, and independence across platforms. As the credit union CRM  had to be utilized by different loan associations in different European countries, we had to face some major differences in the data that had to be processed by the software. For example, in Romania, under specific conditions, the presence of a guarantor is required, and in other countries, it’s optional. That is why Light IT Global’s specialists couldn’t opt for total unification and had to implement separate layers for each organization that was about to use the software.One of the essential parts of the customer evaluation process is running multiple checks (criminal records, credit history, income, social security, etc.) This required seamless integration with different external databases, including government ones. This mechanism is automatically activated once a new loan application is submitted and a chain of checks is launched. Once the data processing is complete, the loan officer receives relevant feedback on whether the customer meets the necessary requirements. Since the customers fill in loan claims on different websites and in different countries, our team also had to develop a component that would transfer all the received data directly to the consolidated system where the managers would be able to process said claims. While most of the approval process is automated, Light IT Global’s experts made sure there’s a way to approve/decline claims and process data manually as well. This change was introduced specifically per the client’s request.As different loan associations’ employees have access to the system, we’ve also incorporated a user role system that defines access rights to various types of data for different users. For example, while processing the claims, loan officers have the ability to upload documents to the system but have no right to delete those. Full control over the assets is available exclusively to a limited number of users. Moreover, the admins can create a multitude of tasks for the staff while each loan officer can see and manage only tasks assigned specifically to them. Such an approach to user management guarantees maximum security for sensitive data and is necessary for industry regulations compliance.

The delivered system helps loan associations in different corners of Europe perform multiple tasks more efficiently. ML-powered software combined with a custom CRM for credit union improves the quality of the credit score evaluation, creates a bespoke loan repayment plan, and helps to analyze diverse data coming from several external sources and related to each particular application. An easy-to-use task manager streamlines the workflow of loan managers and makes it easier to track the performance efficiency of both individuals and teams. The ability to make manual changes to the automatically processed claims safeguards an individual approach to working with clients and provides all the necessary flexibility and adaptability a financial business might require.

Integration of a bespoke ML-based platform and a CRM for credit union helped our client minimize financial losses as the system prevents high-risk transactions from happening. At the same time, the digitization of the workflow allowed the loan officers to process more applications and increase overall productivity. Innovative decision-making software reduces the application validation time, while the “manual approval” option leaves more options for adjustment. As a result, several companies that have already introduced this credit union CRM tool into their workflow managed to increase operational efficiency and boost profitability while simultaneously lowering the risk of fraud.