Cardabel The Alternative 'Inappropriate Trading' Detection Solution

The First Machine Learning Solution to detect known and unknown types of inappropriate trades in Capital Markets


Why are banks still exposed to Rogue Trading?



Cardabel's Detection Engine



Prevent Unexpected Losses

Fraud prevention and detection are very serious challenges for banks. Cardabel's software detects trading strategies that deviate from the expected mandate, abnormal values of transactions & the abnormal evolution of key indicators.


Better Detection &
Easier Analysis

Cardabel has designed, adapted and benchmarked Machine Learning algorithms that are dedicated to this industry & this problem. Not only does it detect known and unknown types of anomalies, Cardabel's software reduces the number of false alerts.


Fast deployment

Cardabel's solution is designed to be easily integrated with bank data. In a matter of weeks, banks can have a background audit of existing banking transactions, which detects anomalies, inappropriate trades, operational losses or frauds.


What does Cardabel detect?


The capability to detect anticipated fraud schemes is essential, but it is not enough. What about unknown fraud schemes? As rules evolve and regulations change, fraud will evolve with them. Cardabel is unique in that its dynamic pattern recognition ensures that no scheme is hardwired in the fraud detection engine. Since it recognizes past schemes without being programed for them, it also recognizes new ones. Cardabel is the only product on the market today that can truly sanitize your operations.
As much as you can prevent a trader from trading a set of currencies or a set of products, it is impossible to create rules for all combinations of values and trade attributes. Even without preset rules, Cardabel detects these errant combinations before they can have an adverse impact on the profitability of the bank.
When P&L is recognized before the maturity of the trades, there is an opportunity to book trades, recognize the P&L (and so the associated bonus) and then amend/cancel the trade after the bonus is paid. This operation can be repeated many times and may be impossible to catch manually. Cardabel exhausts the build-up of patterns amongst trades and trade aggregates and has adapted time series pattern recognition to highlight exceptions from expected patterns.
Kerviel had Eurex positions representing 85% of all positions held by Societe Generale. Aggregating and monitoring his positions and associated ratios would have detected this at its earliest stage. Cardabel generates and checks all meaningful aggregates and ratios and then detects outliers to normal values.

Cardabel's unique advantages



Full coverage: detects
all types of anomalies

Using Unsupervised Machine Learning, Cardabel's software can detect known and unknown types of anomalies, frauds or operational issues.


Optimizes Resources Allocated
to Anomaly Analysis

Cardabel's solution prepares a full documentation of each anomaly in order to reduce the need for skilled analysts.


Easily integrated with
bank data

80% of entreprise implementations are dedicated to integration. Cardabel has avoided this pitfall by directly using the source data without mapping. Cardabel can integrate with any type of data source.

Learn more about the product