HSBC has deployed AI systems to automate cash management for ATMs.

HSBC (NYSE: HSBC) has developed in-house AI-powered technology to fine-tune cash-filling operations at its ATMs, currently outsourced to third-party vendors. The system was successfully tested on the bank’s network of 1,200 ATMs in Hong Kong. (ComputerWeekly.com)

The bank saved as much as $1 million because the number of ATM refills declined by 15% following the use of the “iCash” AI technology.

How HSBC’s iCash works

HSBC has trained machine learning algorithms on cash withdrawal trends factored with seasonality, holidays, public events, and the ATM’s location. These “predictive” algorithms are run with live ATM data. The algorithms calculate the amount of cash needed at each ATM location.

According to HSBC, iCash is a big improvement on its earlier system of manually forecasting cash demands at its ATMs. Because the method was not so accurate, the bank faced customer complaints due to cash drying up at ATMs. However, the solution of repeated refills or larger ATM cash stocks was inefficient.

Moreover, scheduled and repeated cash refills also increased the risks from criminals launching a cash heist.

Under the new system, HSBC personnel have live dashboard visualizers showing cash withdrawal trends. They are thus able to optimize the cash refilling operations on a ‘just-in-time’ basis. iCash has slashed lead times on cash restocking from 36 hours to just 15 minutes.

“iCash is a game-changing digital solution that improves the customer experience while unlocking both man-hour and vendor savings,” said Chris Trill, global head of wealth and personal banking operations at HSBC. “It also reduces the risk of robbery by moving away from scheduled cash deliveries.”

HSBC now intends to deploy the technology at its locations outside of Hong Kong.

However, other AI applications for ATMs already exist.

Predictive maintenance of ATMs and facial recognition of customers

According to Emerj, IBM’s (NYSE: IBM) Technology Support Services as a predictive maintenance software that can identify the likelihood of maintenance problems at ATMs. It can also automatically schedule the necessary maintenance services.

A similar service is offered via the NCR SmartServe Predict solution which can analyze historical sensor data from ATMs, predict any performance problems, and alert the maintenance team. NCR (NYSE: NCR) is one of the largest ATM manufacturers in the world.

Uncanny Vision, a start-up based in Bangalore, India, has developed software that delivers real-time video analytics and surveillance for ATMs. The software can analyze the footage from an ATM surveillance camera using a deep learning-based vision system in real-time located in the camera itself.

Diebold Nixdorf (NYSE: DBD), another very large ATM manufacturer, and Samsung (KRX: 005930) collaborated on the development of a cardless ATM that uses facial recognition combined with NFC technology integrated into the machine. The system achieves two-factor authentication. Users can tap their mobile devices on the machine’s NFC reader and thereafter receive an authentication alert via facial recognition.

Related Story:   PayPal: “We will Hasten the Demise of Cash”