Bank Negara Indonesia taps voice biometrics for password resets

Tired of manually resetting forgotten staff passwords, Bank Negara Indonesia (BNI) has automated the process using biometric voice authentication technology.

The bank has installed Indriya Ease, a multifactor authentication system powered by Nuance Communication's VocalPassword technology.

Supporting the local Bahasa Indonesia language, the system lets BNI's IT help desk hand out password replacements to staff who verify their identity using voice samples.

M Harsono, IT helpdesk and command centre group head, BNI, says: "Manually verifying the staff's identity maybe the most secure option, but it causes inconvenience to both IT helpdesk and users. We felt that a convenient and secure automated password reset mechanism is a must to cater to so many requests."

FInextra verdict Interesting application, but we'd love to see the ROI on this, particularly when taking account of the enrolment procedures for new staff members and the handling of false positive referrals back to the help desk.

Channels

Retail banking Security

Keywords

Authentication Biometrics Voice recognition/speech verification

Comments: (1)

A Finextra member 29 March, 2012, 08:13

Be the first to give this comment the thumbs up

0 likes

Making the business case for biometrics with a 'closed' group of employees no doubt easier than for an 'open' group of customers.

Costs for a biometrics solution fall under a number of headings - capital investment in the system, cost per registration (to create the reference biometric) and cost of validation (checking identity).

With employees, the bank can at least insist on registration, with is much more difficult with a customer group who don't necessarily see the value.

But the other question remains - why is the bank applying a higher degree of security just because the system is automated? When an employee calls the help desk, what checks does the agent apply? Most likely, they just ask for some personal details - an employee number, maybe a department name or email address, maybe date of birth. All of these could be handled by straightforward speech recognition, no need for expensive biometrics and no need for pre-registration.

So even if there is a business case for biometrics, there is a better one for a simpler solution ...