Inside Mastercard’s AI Architecture

Advertisement

“There are some things money can’t buy. For everything else there’s Mastercard” 

But what makes Mastercard tick? The short answer: AI. The global payment & technology company leverages artificial intelligence to enhance data safety and reduce cyber risks. After its initial deployment of AI and ML in 2016, Mastercard has acquired Brighterion and Ekata to strengthen its AI capabilities.

THE BELAMY

Sign up for your weekly dose of what's up in emerging technology.

Mastercard’s AI strategy is built on “five pillars”.

Source: Brighterion

“Our position is that this technology is the next electricity, and we want to deploy it at scale across Mastercard and ensure our customers can benefit from the same,” Rohit Chauhan, executive vice president at Cyber and Intelligence Solutions of Mastercard, said. Mastercard also has its AI consulting service (AI Express) to help businesses develop their own ML models. AI creates Mastercard’s financial forecasts 90 days in advance to predict the volume of transactions coming up. 

Mastercard aims to use AI as a tool to advance equitable economic development in developing countries to enter a global economy where they have access to financial systems. This can be done by providing access to financial instruments like credit and working capital.

“As easy as it is for the consumer, the complexity lies in the background — we have seen the evolution of this hyper-connected world in the backend just explode,” said  Johan Gerber, executive VP for security and cyber innovation at Mastercard.

Fraud detection and prevention

Mastercard deployed AI solutions to combat fraud, money laundering and credit risk 20 years ago. Today, AI is involved in every customer interaction. Mastercard has developed an in-house solution called Decision intelligence applying thousands of data points and modelling techniques to each transaction to approve genuine users in real time. AI checks data points including location, merchant, device data, customer value segmentation and risk profiling. The application will then check the merchant’s system to see if the user has a risk or trust score. It will approve the transaction if it finds no consistent patterns of fraud. False declines and fraud detection were among Mastercard’s main concerns while developing its AI model. 

AI’s ability to self learn brings about greater accuracy as opposed to rule-based fraud detection systems and allows personalisation of fraud detection. Traditional detection systems that use predefined data like location and so on to group customers and cannot fully grasp individual behaviour. On the other hand, AI mines historic and real-time data from wider touchpoints identifying changes in an individual’s buying behaviour.

Biometrics 

Mastercard’s Identity Check uses machine vision to identify customers using biometric data. With the Identity Check feature, a fingerprint or eye blinking is enough to authenticate the payment. The solution is powered by EMVCo’s 3 D Secure authentication messaging protocol. The Three-Domain Secure enables customers to make payments and purchases without a card. This provides an additional layer of security to the three financial domains involved in a transaction- the merchant, issuers and the payments systems moving customer’s money. 

Building blocks

Companies don’t usually build their technology from scratch. Mastercard recently acquired the identity verification startup Ekata to improve the customer experience. 

In cybersecurity, isolation may minimise the possibility of data leaks and cyberattacks, but most systems make use of “lego blocks”, as Gerber puts it, to share intelligence within the industry (systemic risk). Collective systems can be put on alert if a cyber attack takes place, which is Mastercard’s philosophy of Connected Intelligence.

Platform design

Systems derive insights from Mastercard’s data and the data is safeguarded with a platform approach. The bottom end is where the raw data is fed, and the company uses various tools and technologies to process numerous sources of data in real time. ‘Intelligence blocks’ are created from this raw data, and this data is used to build AI models and products. No data is shared with third parties like banks or retailers, but Mastercard can still approve transactions on an individual level. This data also gives the company important insights into cyber threats with a lot of automation at play.

More Great AIM Stories

Prajaktha Gurung
I am a literature, media and psychology grad which explains much of my confusion in life. I like writing, especially about music. You'll find me clicking photographs and playing music on my guitar most of the time!

Our Upcoming Events

Conference, in-person (Bangalore)
MachineCon 2022
24th Jun

Conference, Virtual
Deep Learning DevCon 2022
30th Jul

Conference, in-person (Bangalore)
Cypher 2022
21-23rd Sep

3 Ways to Join our Community

Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

Telegram Channel

Discover special offers, top stories, upcoming events, and more.

Subscribe to our newsletter

Get the latest updates from AIM
MORE FROM AIM