Making selfies pay

“Selfies to pay” could emerge as a new lure to bring younger customers into the banking fold as face recognition matures as a widespread technology within the financial sector, writes Roland Tellzen

By Jon Watkins

Speaking to Sibos 2018’s Discover Zone, the founder and chief executive of Amsterdam-based VisionLabs, Alexandr Khanin, outlined how Russia’s Sberbank has implemented the company’s facial recognition system across its network. The Sberbank application, with some 110 million faces on its database, is the second biggest deployment of a national-scale facial identity system in the world.

Implemented in 2017, the system is in operation across 18,000 branches in Sberbank’s network, as well as available online via connected and mobile devices.

“It is not a prototype, it is a working product,” Khanin said. The Sberbank system processes more than 1500 recognitions per second, with an allowed latency of less than one second to process each. It’s engineered to ensure an accuracy of less than one false-positive ID for every million searches, with each customer’s facial ID data stored within 250 bytes.

Khanin said while the system has obvious benefits in terms of security and access to accounts, its value as a marketing tool also has been proved. “There is huge competition between banks to retain existing clients and to attract a new audience,” he said. One effective way to do this is to “digitise your audience”, he told delegates.

A selfie to pay would link users’ faces to their wallets, phones and bank accounts. A user would just have to “blink or smile” to prove a transaction using the system, said Khanin. “It is much easier and much more fun to use your selfie, and this can help get a new audience of younger people.”

For marketing departments, the technology can provide a detailed and concise snapshot of the bank’s day-to-day operations and users. “Using facial ID, you can count your total number of visitors to individual ATMs and branches. This can then be further refined to unique visitors, as well as such segments as age groups, gender, retention rate and such.

“This snapshot can be used to predict and prepare for expected traffic flows in individual branches at different times. Even further, it can help banks to better serve their very important customers by recognising them in real time. It doesn’t matter what branch they are going into, they can be made to feel at home.”