Machine learning (ML) sits at the heart of advanced fraud prevention platforms like Simility, a PayPal Service. But there’s a difference between boasting of the technology on a specs sheet and delivering it in a meaningful way for users. Data scientists and in-house fraud teams are in short supply today, making it vital that organizations maximize their productivity, and their ability to drive value. This is where integrated dashboards can be a crucial resource: offering a single pane view to help teams answer the questions they need to in order to drive long-term business success.

Simility’s ML Dashboards feature will help customers achieve exactly this.

The Power of Machine Learning

Fraud is constantly evolving, and as it does so patterns are becoming increasingly complex. Today’s scammers operate across geographical boundaries with anonymizing tools to hide their true identity, and access to large volumes of breached logins and identity data to impersonate legitimate customers and invent new ones. They’re even using machine learning and automated botnet-driven tools to improve success rates and outwit legacy fraud prevention tools.

Tackling this threat demands a sophisticated response, through ML-powered tools that also leverage large volumes of structured and unstructured data — from inside the business and a range of third-party sources. Simility’s approach uses ML models to make ongoing recommendations to fraud rules, suggesting new thresholds and then simulating rule performance by comparing the changes to historic model data. This automatic tuning enables businesses to improve the accuracy of their decisions whilst freeing-up fraud teams for higher value tasks and staying ahead of scammers.

Time to Visualize

However, while much of this work is automated, there’s still a crucial role for fraud analysts and data scientists to play — leveraging the power of ML tools to make better informed decisions and spot emerging fraud patterns. This is why data visualization is a big part of Simility’s platform. Our UI was designed to display rules, ML, and Device Recon data in a single analytics view, which users can manipulate to slice-and-dice, drill down and identify connections between data points to uncover fraud patterns.

Dashboards are a natural extension of these capabilities, providing a one-stop-shop for customized data views that provide high-level intelligence at-a-glance. They sit at the heart of Simility’s efforts to maximize ML transparency, boost user productivity, and leverage ML to optimize decision making.

Introducing ML Dashboards

That’s why we’ve introduced a new dashboards feature in the Simility platform, designed to help customers answer key questions about their ML models and their impact on the business of fraud prevention.

With ML Dashboards, users will be able to view a simple report for each of their ML models. This includes: current and realized fraud; financial losses; the number of chargebacks and declines; confusion metrics; and the automation rate, among other details. With this information, data scientists and fraud analysts can quickly answer questions such as:

• How well is my business doing?
• What is the current and closing business for the last claims window?
• How much business have I lost?
• How much business have I saved?

With this kind of rapid insight, users can make better informed, more strategic decisions to drive business success and improve product performance. As fraud challenges escalate, it’s the kind of insight into ML models that businesses need today.

To find out more about ML Dashboards in the Simility platform, schedule a demo today at simility.com/demo