EQUIFAX ANNOUNCES IMPRESSIVE RESULTS USING
NEURISTICS CHECK AUTHORIZATION RISK MODEL
NewsEdge Corporation has reported that Equifax, a leader in consumer and business information, and Neuristics Corporation, a leading provider of intelligent decision-support tools, have announced the first year's results of the implementation of an advanced artificial-intelligence (AI) model to reduce risk in check authorizations. Through use of the model, Equifax's retail customers have realized increased sales through greater approval rates and achieved increased efficiencies in the check authorization process.
The AI model was in development and testing for more than a year, using millions of transactions and historical data sets to develop, benchmark, test and validate the model.
"Responsible for the authorization of billions of dollars of checks each year, Equifax needs the most powerful and effective means available to determine whether or not the funds will be available to pay a check," said Andrew Krause, Neuristics CEO. "The team effort of Neuristics and Equifax to implement an intelligent decision-support solution has resulted in a better, more effective check authorization process."
Equifax Check Services incorporated the AI model into their check authorization risk management platform -- the Expert Authorization System (EASy) -- in mid-1997, pioneering the use of neural network technology in combating the rising costs of handling high-risk transactions and improving the effectiveness of its decisioning process. Results since its implementation have shown a 17% increase in the dollar value of approved check transactions (increased sales to the retailer), with a corresponding 3% reduction in the dollar value of returned checks, and a 40% gain in check collectibility rates. In addition to increased sales and lower loss rates, a 15% reduction in processing time has created a quicker, more efficient check approval process for the retailer and customer at the point-of-sale.
The AI model -- called the Check Authorization Risk Model (ChARM) by its developer, Neuristics Corporation -- predicts whether or not the check will be honored, either initially or during the loss recovery process. To make this determination, ChARM evaluates information about the check writer, the current check, and the circumstances surrounding recent check writing activity, including an assessment of risk factors for recently presented checks throughout the entire Equifax client base.
In addition to cost savings and check authorization efficiency, implementing the ChARM model provides Equifax with a new, highly sophisticated set of decisioning tools to fight check writing fraud. "When we wanted to cut costs and improve the speed of our check authorization process, we turned to a company we could trust to get the job done.
"Meeting an aggressive development schedule, Neuristics provided us with a solution that is proving extremely effective," said Jeff Carbiener, Senior Vice President and General Manager of Equifax Check Services." Basically, including the model in our decision process is allowing us to fight the new wave of 'high-tech' fraud with innovative 'high-tech' solutions."
Equifax serves the banking, financial, retail, credit card, telecommunications/utilities, transportation and healthcare industries and government. Founded in 1899, Atlanta-based Equifax today has 13,000 employees in 18 countries and sales in more than 40 countries. Revenues for the 12 months ended June 30, 1998, were nearly $1.5 billion. For more Equifax information, visit the company's website at http://www.equifax.com
Neuristics Corporation develops and markets intelligent decision-support tools that enable financial services providers to uncover new opportunities for profit and to reduce risk and overhead in acquisitions and relationship management in consumer credit and payment systems. Neuristics clients include major banks, financial transaction processors, and a third of the top 25 credit card issuers. Founded in 1993, Neuristics is a privately held firm based in Baltimore, Maryland.