App for Getting Instant Loans / Online Lending Platform for Small Businesses

Project Background

The client is a FinTech startup with decades of experience in the financial services industry. Recognizing many inefficiencies in the current loan business they decided to launch a fully digital online loan platform and a mobile app for small and midsize business that would fully automate traditional loan business providing the following benefits:

• Allowing the end clients to apply for and get a loan and make payments via a mobile app in minutes not leaving their home
• Lower operational costs for capital providers and lower the interest rates for the end clients through full automation of the process and minimizing human involvement
• Allow disbursing more loans with a lower default rate with AI-based self-learning credit scoring module
• Move operations from brick-and-mortar branches to the online platform

The client was looking for a technical partner with profound expertise in the Fintech industry, namely digital lending technologies, artificial intelligence, and mobile app development. Itexus was selected for its expertise in those areas and for its flexible startup-oriented approach. 

Functionality Overview

• A mobile app for the end clients with user registration, KYC, loan application, agreement signing via DocuSign, virtual credit card issuance, payments, statistics and reminders functionality.  
• Administration Module with overall stats of app performance, user management, scoring settings, and reporting.
• Back office with advanced reporting and loan portfolio monitoring functionality. 
• External Integrations with 
  • Plaid – bank account aggregator
  • Experian – Credit history bureau
  • Marqeta – Virtual credit card service from Visa
  • DocuSign – Agreement signing
  • Stripe
• Advanced credit scoring model using credit history and transaction data and an ensemble of statistical and machine learning algorithms to determine credit risk, interest rate and other parameters. 
• Automated Know Your Customer (KYC), Anti Money Laundering (AML) processes through integration of the industry’s leading KYC/AML providers such as Experian.
• Automated bad deal management module. Automatically selling nonperforming loans to a collection agency.

Project Approach

Discovery phase

The project started with a discovery phase, during which  the Itexus Business Analyst and Software Architect team performed and In-depth market and requirements analysis and created the initial project documentation:

Software Requirements Specification

Document describing all functional requirements with use cases, diagrams, user screen mockups, user journey etc.

Software Architecture Document

Document describing suggested technology and architecture of the system addressing, third party integrations, security, performance, reliability and other non-functional requirements.

Project plan and a detailed work estimate

Detailed project plan with all work broken down into 8-16 hours tasks, with priorities, dependencies, and team structure.

UX/UI Design

At the start of the project our team of UX/UI specialists designed a user-friendly intuitive UI of the mobile app. The UI mockups were combined in a clickable prototype and a marketing video that were used for marketing purposes long before the system was ready.

The delivered design is based on the Apple Human Interface Guidelines.

Development

• Agile/Scrum development process with 2-week sprints and a demonstration of the new product versions and feedback collections session at the end of each sprint
• Continuous integration and deployment process
• Combination of unit test, automated service and UI level tests and manual testing 

Artificial Intelligence Based Credit Scoring

A dedicated team of data scientists on our side worked in close collaboration with credit bureau specialists to create an AI-based credit scoring module that used credit history reports, transactional and social data on both the business and the business owner, assessing the value of the collateral, future inflation predictions, and overall economic growth to forecast the probability of default on a loan and calculate the optimal loan parameters in real time.

The credit scoring module used an ensemble of algorithms varying from logistics regression to deep neural networks to achieve optimal performance on any volumes of data.

The models are updated and retrained on daily basis as the new data comes in. 

Results & Future Plans

The final product has been delivered within budget and on schedule, ready for launch in the App Store.

The client is currently negotiating deals with major U.S. and local community banks to launch the financial platform as a means to deploy capital through the platform.

Itexus team is working on the second version of the product turning it into a white label solution.