Most major financial institutions have a mobile app. These applications are customer gateways into banking services, such as moving and managing money. Application usage already sits at 800 million per day, and the number doing most their financial dealings through mobile will increase. Preventing fraud is difficult, especially on platforms like Android where security is less than stellar. Microsoft has published a blog post that shows how the company has leveraged AI within Azure to combat mobile fraud. The model can detect fraudulent situations in less than two seconds, giving customers time to prevent attacks. Microsoft says most observed mobile fraud situations are achieved through a “SIM swap attack”. This essentially occurs when a mobile number is either hacked or cloned. The bad actor then receives all messages and calls for that number. Microsoft’s AI model works within Azure creates a behavioural overview and detect fraud attempts more rapidly. “Latency and response times are critical in a fraud detection solution. The time it takes a bank to react to a fraudulent transaction translates directly to how much financial loss can be prevented. The sooner the detection takes place, the less the financial loss.”
Leveraging Azure Services
Microsoft used three separate components to underpin the new AI, all of them based in Azure. The company used Azure Functions, Azure SQL, and Azure Machine Learning to create the following three factors:
Feature engineering to create customer and account profiles. Azure Machine Learning to create a fraud classification model. Azure PaaS services for real-time event processing and end-to-end workflow.
To help users understand more about mobile banking fraud and ways to prevent it, Microsoft has published a Mobile bank fraud solution guide. This document also details the modelling behind the AI and what it could achieve.
