The Caspian Data Diagnostic tool makes it fast and easy to analyse the quality and integrity of your data. Our labelling and training pipeline then tailors the platform’s models to your data sources and risk thresholds.
Our Platform Management tools then empower you to analyse the performance of the AML Investigator platform prior to going live and apply any corrective action required.
In order to configure our models, Caspian needs access to your historic case data.
These can be supplemented with data from your propriety or 3rd party sources to enhance the platform we provide to you.
The Caspian Data Diagnostic tool makes it fast and easy to ensure the quality and integrity of your data is suitable for training and labelling purposes.
The models are trained to match your policies, processes and risk profiles, based on your experts labelling of a sample of historic case data.
Our Labelling Tool makes this process rapid and consistent, with multiple experts’ labelling the same transactions so that we can apply consensus methods to establish a true gold standard.
The aim of the labelling process is to link specific pieces of data to the defining features of a transaction.
Our pre-configured models are then trained on your experts’ consensus. The process is automated, but you remain in control over which version of the model is promoted or reverted. A complete audit trail ensures that there is always a complete record of the evidence behind every single decision.
Each version of the trained models are analysed before going live to ensure that they meet your risk thresholds. Performance is analysed at a granular level – including for each type of alert – because aggregate error rates can disguise issues with rare, harder to detect, but high-impact types of case.
This ensures that performance is above threshold for each type of customer and alert.
We are delivering the Caspian AML Investigator platform in Tier 1 Banks to transform high-volume investigation management. We investigate and judge a high volume of AML alerts more efficiently, accurately and to a scale that exceeds in-house experts.
REDUCED OPERATIONAL COSTS