OUR USE CASE JOURNEY
A financial services company wants to optimize its working capital and reduce over or under-supply margins across branches, where it performs money transfer services.
Over/Under Supplying. All the company’s branches were suffering an over or under supply of 300% when restocking. Consequently, some branches closed early and some were paying too much insurance.
Limited Capabilities. Strict regulations were enforced from the Central Bank of Egypt limiting the company’s services to 40 branches and no money transfer abroad.
Poor Cash Management. The company had an unoptimized working capital amount in addition to a high debt/equity ratio.
To capture patterns in the data, a multitude of relevant data sources were considered to reach the most accurate results.
Data Infrastructure. A data cleansing pipeline was implemented to prepare the data before being consumed by the machine learning solution.
Mixed Data Sources. In addition to the main data sources, external data was taken into consideration such as transactional data, macroeconomic data and circumstantial data (e.g. weather conditions).
The Algorithm. Instead of using the classical forecasting techniques that led to the aforementioned errors, a blend of algorithms was created to come up with the appropriate equation to govern all forecasting needs with extremely high accuracies.
Working Capital. The platform saved more than 15% of working capital after implementation.
Over/Under Supplying. The over/under supplying error decreased from 300% to 12%, saving millions of dollars monthly.
Continuous Learning. The algorithm learns and gets better with time; after deployment, the error decreased from 15% to 12% after a few months.
The project is successfully running, minimizing inaccuracies and increasing scoring quality as it enters phase 2: The Insight-Driven Transformation phase.