Case studies

Analysing financial transactions at unprecedented speed and accuracy

Fintech Open Banking Company
South Africa

Our client is a fintech provider at the forefront of the open-banking movement in South Africa.

Their customers, typically banks and other financial services providers, utilize the platform

to access customers’ banking information in a secure and reliable way. Additionally, the platform

assists financial services providers through superior analytics capabilities to gain insight into

banking transactions thus empowering their customers to make smarter decisions.

The Problem

The client's customers regularly conduct affordability and credit assessments on their customers based

on banking transactions. A major component of these assessments is identifying an applicant’s

salary accurately in order to understand the composition and regularity of income.

Bank transactions are unique to each individual and descriptions contain text which is difficult to

analyse using traditional statistical techniques. Therefore, trained staff had to manually identify

these transactions in banking statements. This process is time consuming and rife with errors even

with highly trained staff.

The end result is that customers have to wait longer to get feedback resulting in poorer service.

Moreover, the variability in the process may lead to less accurate assessments leading either to

potential losses by the company or unnecessarily limiting customers’ opportunities for financial


Our Solution

Elucidate developed a series of models to identify salary items in bank transactions involving

cutting edge Natural Language Processing (NLP) techniques in order to understand text descriptions

and consider consumer’s bank transactions more holistically. The system learns and improves over

time and is becoming more effective as it sees more transactions. The system is able to respond in

real-time and provide reliable results far surpassing human performance.

The Results

Clients are able to automate salary identification in affordability and credit assessments. At

over 99% accuracy, this is done far more accurately than in the past. This results in superior and

faster service to customers, significant time and cost savings for the financial service provider and

leads to smarter more accurate assessments.

1. Over 99% accuracy

2. Surpass human expert performance

3. Less than 0.8 seconds response time

4. Automate processes previously considered not automatable

Similar case studies
Huduma Credit

Huduma Credit is a lending company based in Nairobi. They provide loans to taxi owners through an efficient, automated digital platform.

Read More
Fintech Open Banking Company

Identify salary items in bank transactions using cutting edge Natural Language Processing (NLP) techniques

Read More
Marketing Services Provider

Automated data intergation and reporting platform to facilitate real-time reporting across the organization.

Read More
The Opus UI Kit lets you showcase your work in style. It's helped take our business online
Alexa F.

Designer @stamps

Not able to tell you how happy I am with Opus. Best. Product. Ever! Wow what great service, I love it!
Alexa F.

Designer @stamps

Let's Talk!

Become Our Next Success Story

Let's Talk