ReD's suite of PRISM® risk management solutions, provides predictive fraud detection and case management for credit, debit, retail and commercial card fraud, as well as merchant fraud, money laundering and e-commerce fraud. These products are helping some of the largest financial institutions and retailers, in the world control fraud losses and reduce the threat of illicit funds moving through their institutions.

The secret behind these intelligent solutions is patented neural network technology. Neural networks are a form of artificial intelligence that function much in the same way as your brain. Unlike rules-based exception reporting systems, neural networks learn continuously from data.

And, where rules-based systems are efficient at detecting known fraudulent activity and criminal schemes that can be cast in a rule or table-entry, they are, nonetheless, static. Predictive neural networks are the perfect complement, providing an adaptive and early warning system for new and ever-changing criminal tactics and customer behavior that have yet to impact profitability.

ReD's PRISM neural network solutions are truly unique, dynamically learning changes in your business and markets. For example, bankcard fraud and money laundering schemes are transient behaviors that can change as fast as existing schemes are uncovered. Rules-based fraud detection applications and systems that implement conventional neural network models are limited, deteriorating over time as the behavior patterns impacting your business change. ReD's PRISM is designed to keep pace with this volatility, enabling you to update the predictive neural models with current transaction data, customer demographics and criminal tactics - at your facility, according to your schedule and with no interruption in system processing.

Retail Decisions is an experienced partner on Risk Management that works with you to gather the data sources that define your customers' activities and fraud experience, installing the products and provide complete, on-site training that addresses system use, interpretation of reports and the techniques involved in using PRISM to optimize, through model updating, existing neural network models.