CubX will change your factoring business.

CubX is a solution that uses state-of-the-art machine learning algorithms to analyze data, including credit scores, financial statements, and customer payment history.

CubX uses data from today and in near real time, allowing us to accurately predict the likelihood of late payments or even defaults.

CubX is made to identify risks early and classify them based on the creditworthiness and customer behavior of debtors, clients and receivables. It detects and predicts chances and risks in the ecosystems of a factoring organization.

Benefits of CubX

in a nutshell:

Scalability

CubX is scalable and thanks to its state-of-the art machine learning technology, it adapts to your particular ecosystem in a short period of time.

Potential opportunities

CubX enables our clients to identify not only various risks but also potential opportunities across clients, debtors, industries, or regions, allowing them to do so at an early stage within the respective ecosystem of each factor. Utilizing the full potential of factoring data.

Timely warnings

With CubX you can receive timely warnings about possible payment delays.

Risk monitoring

CubX offers you an improved portfolio for risk monitoring.

How it came about

Our systems generate vast amounts of data annually—150 to 200 million invoices and a transaction volume of €200 billion. But what are we doing with this data beyond simply processing it? How can we further analyze and leverage it?

To explore this, we consulted our data scientists to assess the potential of the data and determine if it could be used for predictions and forecasts. The data sets we used included client master data, debtor master data, and invoice payment data, sourced from ef3 and efX.

We implemented a comparison between two approaches:

A traditional expert-driven, rule-based method, where we engaged both internal experts and customer insights to analyze the data.
A supervised machine learning-based method, where we input the data into the system and evaluated the results generated by the model.

It became clear that the machine learning approach provided several advantages over traditional data processing techniques.

Among these, machine learning allowed us to process vast amounts of data rapidly and identify complex patterns and relationships. Furthermore, and most importantly, it enabled us to make data-driven, objective decisions, rather than relying on human judgment, which is often influenced by noise and biases. These are just a few of the many advantages the machine learning approach offers.

Recognizing these advantages, we explored further applications of the machine learning approach to analyze data such as credit scores, financial statements, and customer payment history. This led to the development of CubX, a tool designed to identify risks and opportunities early on.

How it works

Data collection

Clients’ data from at least one year is collected (often from the oracle databases) and anonymized.

A select team of internal data scientists have access to this data.

The data is extracted and saved to our efcom oracle database to be worked on with the SQL Quary output system.

Storage

To keep large amounts of Data updated and on a real time basis it needs to be scaled and maintained. That’s why we use the OCI (Oracle cloud) with oracle buckets. This allows us to keep the data scalable, aways available and process Data fast.

To run complex algorithms, a special platform is needed. That is why we work with “OCI Notebooks”, which is also equipped with an Identity access management system.

Application Access

A user-friendly dashboard with clear visualizations, along with an integrated Identity & Access Management system, is implemented to ensure data protection.

Lastly, access is granted through a firewall-protected system for verification purposes.

Conventional vs. ML risk management

Without ML

Without machine learning risk management is a manual, rule-based process, where risk is identified through predefined criteria or ratios. Professionals analyze data to assess risks and develop strategies accordingly.
Financial statements, credit agency scores, payment behavior analysis, as well as industry and market trend assessments, are commonly used in this process.
As the volume of data grows, the challenge of processing large datasets efficiently increases. This can lead to inaccuracies in the analysis.
Additionally, this traditional approach often reacts to risks after they occur, rather than proactively predicting and preventing them.

Machine Learning

Machine learning can analyze and process vast amounts of data at high speed.
It allows the identification of complex patterns and relationships within data.
With machine learning it is possible to adapt to quick market changes.
Additionally, it enables automatization of repetitive or time-consuming tasks.
By relying on data-driven algorithms rather than human judgment, machine learning can reduce personal biases and subjective decision-making, leading to more objective and consistent outcomes.