November 08, 2017
Is your fraud prevention system a black box? It’s time to take control
Whenever we speak to fraud and risk management teams around the world, we hear that they would like to move away from constraining “black box systems” offering little to no influence on model configuration and other settings.
There is an increasing desire to combine the advanced machine learning capabilities of third-party platforms with in-house advanced analytics. This allows organisations to:
- Manage their own data science models and rules within an external solution – for even more powerful data analysis
- Get the detailed, real-time insight into fraud events that they need to have full confidence in advanced Artificial Intelligence and machine learning technologies
From black box to Sandbox – what can ARIC™ 3.10 do for you?
With the new ARIC 3.10 release now available, we have added even more functionality to enable organisations to take control of their fraud defences. The 3.10 release offers you the award-winning real-time machine learning capabilities ARIC is renowned for, with even more insight into your data and analytics.
Features such as Sandbox Replay as well as security and UI enhancements, improve the functionality and continued management of ARIC while taking into account future growth and development. Sandbox Replay is a user-friendly test arena to check the impact of your changes to the system’s performance.
What are the new key features in ARIC 3.10?
Fitting seamlessly into the ARIC Analytics Configuration Workflow as part of the same user interface, Sandbox Replay enables users to test the performance impact of changes to the analytics without affecting the live system. This means you can compare how ARIC would have performed with new or revised rules, models, aggregators, risk thresholds and other analytics versus its actual performance over the selected time period. The testing can be completed within minutes, allowing you to test and tweak much quicker than before.
Fig 1: Sandbox Replay fits seamlessly into ARIC’s Analytics Configuration Workflow.
Active-Active Disaster Recovery
With Active-Active Disaster Recovery, ARIC 3.10 can be deployed with service replicas distributed across multiple data centres, providing disaster recovery in the event of a catastrophic failure at one site. In the case of this happening, ARIC can continue normal event-processing operations without manual intervention, providing resiliency against temporary or permanent interruptions of data centre operations.
User Interface (UI) Customisation
The new UI Customisation functionality enables you to edit the layout and content of a number of key incident management screens and add data from external sources. In addition, users can format and customise fields, including using country flags for events with geolocation information.
Data Connector Management UI
With the Data Connector Management UI, you can configure how ARIC ingests batch data and data from databases, files or message queues. You can also validate a new or changed configuration before saving to make sure the settings do not affect live event processing.
Enhancements to messaging service and model retrains
The new release also includes improvements to ARIC’s built-in user-to-user messaging service as well as model retrain enhancements that ensure minimal impact on live event processing.
Sean Neary, Featurespace’s Subject Matter Expert for Financial Services, comments: “The 3.10 release adds valuable features to the ARIC Fraud Hub. Sandbox Replay is very important and exciting as it enables our customers to test their data science models and assess their performance in a safe space where the live system is not affected. With the added reliability, security, and customisation benefits that come with the other features, 3.10 will really help our customers to understand and take control of their fraud management system.”
Want to find out more? Speak to your account manager to schedule your specific release if you already are a Featurespace customer or select from the following options: