Customer Story

Banking and Financial Services

Fraud Detection

Want to learn more?

Contact us

Background

One of Pecan’s customers is a private bank, which offers personal and business banking, as well as insurance and corporate financing. The bank is confronted daily with sophisticated fraud attempts and lacks technological depth to confront this challenge.

The bank’s security division did not have available data researchers to embark on a data science project that entails training and maintaining multiple deep-learning predictive analytics models. They also did not have time to waste, they needed to understand where issues may arise, and improve their fraud prevention initiative.

Artificial Intelligence, Big Data, Predictive Analytics and other technologies provide financial institutions new business horizons, yet they serve fraudsters as well as means for initiating ever growing sophisticated fraudulent attacks against banks. There fraudsters tend to use: identity theft, account takeover, cyber-attacks, card not present fraud and authorized push payments scams. Fraudsters and hackers are inflicting damage estimated at $5.1 trillion over the last 20 years. A KPMG survey highlighted the specific challenges facing global banking fraud. Among other findings, globally over 50% of respondents experienced an increase in fraud value, and over 60% of respondents experienced an increase in fraud volume.  

 

The Solution

The Company chose Pecan’s automated deep learning predictive analytics platform to address this challenge due to Pecan’s fast time-to-model and significantly lower cost.

The bank’s BI team used Pecan to build a predictive model that signals out transactions that have a high likelihood of being fraudulent. The accuracy of Pecan's prediction reached 81% true positives and 4% false positives. The model was built and tested within a week of access to the raw data (compared to 5-7 months required for alternative solutions).

The Results

Out of a total weekly 7,359 frauds identified using Pecan, the bank was able to block in real time 4,511 attempts. The impact on the bank’s bottom line from this project was estimated at $2.75 million for the test period alone. Compared with the previous fraud detection systems employed by the bank, Pecan increased the fraud detection rate by 50% and at a fraction of the cost and time.


It's time to plug your organization into its future.

Try Pecan