Screening
Screening System for Clients or Prospects of a Financial Institution
ABOUT THE CLIENT
Microfin is a well-established financial institution in the Uruguayan market with over 15 years of experience.
ABOUT THE PROJECT
Industry
FINANCE
Business Model
FIXED PRICE
Client Location
MONTEVIDEO, URUGUAY
Tech Stack
Project Duration
3 months
Team Size
5
HIGHLIGHTS
ENHANCED SCREENING EFFICIENCY
The implemented screening system helped the financial institution perform periodic screening of its client portfolio and prospects in record time.
ADVANCED SEARCH CAPABILITIES
The system allows configuring searches in different list groups, with configurable complexity levels, enabling searches by document, name, or company name using fuzzy or phonetic search techniques.
FULL TRACEABILITY FOR COMPLIANCE
With built-in traceability in lists, rules, and executions, the financial institution has the assurance of complete screening records, which are highly valuable for internal use or regulatory reporting.
CLIENT GOALS
The financial institution aimed to implement a screening system to enhance security and compliance in different use cases:
•
Perform periodic screening of its client portfolio, covering more than 40,000 clients, to detect whether any are listed on local or international blacklists.
•
Automate online screenings for real-time decision-making in processes like credit applications.
•
Enable manual screenings on demand to cover specific use cases as needed.
•
Ensure full traceability of all screenings performed for internal audits and regulatory reporting.
THE STRATEGY
Initial Analysis
After meetings with various stakeholders, an analysis was conducted of the local and international lists to be used, totaling between 500k and 1M records, with the potential for future growth.
Research
Solutions were evaluated, ranging from custom implementations in Java or Python using NLP libraries to search engines like Solr or Elasticsearch. Elasticsearch proved to be the best option.
Proof of Concept
A proof of concept was carried out using Elasticsearch with multiple nodes, loading more than 1 million records, performing search tests with various configurations, and measuring response times.
Solution Development
After validating the proof of concept, the system was developed and integrated with Elasticsearch, ensuring high configurability and traceability for auditing. Periodic meetings and deployments in test environments were conducted.
Production Release
After development and testing were completed, the system was deployed to production with full support for the client.
TECHNOLOGY STACK
Frontend:
REACT
Backend:
NodeJS
Database:
MongoDB
Search Engine:
ELASTICSEARCH
SCM:
GIT & GITHUB
THE OUTCOMES
The Screening system fully meets the client's expectations, performing online screenings in less than 1 second over hundreds of thousands of records. This allows seamless integration into their credit application process, as well as screening their entire client portfolio (40,000 records) in under 30 minutes.
Regarding auditing, which was another key requirement, the system allows viewing all executed screenings, including which lists were used, their respective versions, the configuration at the time of execution, and the user who performed the screening. Additionally, a dedicated auditor profile is available for analyzing this information.
The system also enables the periodic updating of lists as different regulatory bodies release updates.
Some of its key technical features include:
•
Fuzzy Search
•
Phonetic Search
•
List and Configuration Versioning
•
Hierarchical configurations (global, list group, individual lists) for search types and fields to be used
Screening was released in 2018