Voice Biometrics Holds the Key to Improved Customer Experience



Despite the multiple channels, including apps, chat, and email, that consumers now use to contact companies, phone calls are still the number one way companies engage directly with their customers. But many aspects of phone interactions are not ideal. There are frequent difficulties understanding caller intent or accents and dealing with background noises, for example. One of the biggest pain points, both in terms of customer inconvenience and company inefficiency, is call initiation, when customers must identify themselves to be granted access to personal data. Asking for detailed information, such as social security numbers, birthdates, or even cumbersome passwords or PINS, is inefficient and annoying.

New solutions for this problem are emerging. One of the most promising is voice biometrics, a simple, convenient, and secure method of authenticating speaker identities using just a few words. This is possible because everyone has different physical and behavioral characteristics that influence the sounds of their voices. Like fingerprints, these are unique to every individual. Voice biometrics solutions are quicker, easier, and safer than other authentication methods.

They're quicker, because only a few words are needed to compare with the enrolled voiceprint, versus multiple questions and responses. They're easier, because customers don't have to remember obscure, complex passwords. They're safer, because customers are using their unique voices to authenticate their identities rather than pseudo-secret questions that can easily be extracted by social engineering.

Voice biometrics, therefore, provides a rare win-win-win opportunity for customer service operations looking to improve their customer experiences while reducing costs and increasing security. Yet overhauling customer service is never an easy decision or task, especially when it affects both employees and customers. Voice biometrics can impact the bottom line as much as it improves customer experiences.

Why Now?

While the science has been proven for some time and technology maturity was inevitable, one of the biggest hurdles to voice biometrics adoption has been far more cultural—a fear on the part of companies that customers would not accept this technology, or at least not at scale. However, this has now been disproven by a constantly increasing list of early adopters seeing customer acceptance rates ranging from 80 percent to 90 percent.

More pressing is that contact center fraud globally has increased by 45 percent in the past three years. This is particularly true for industries experiencing a high volume of repeat callers, as approximately 30 percent to 50 percent of all fraud incidents are initiated with a phone call. Financial services reported more than 1 million incidents of financial fraud in the first half of 2016, or one incident every 15 seconds, representing a 53 percent increase over the same period last year. Industries with similar risk include telecommunications, government, health care, insurance, and digital commerce.

Often these attacks stem from social engineering or data breaches outside the organization's control, which highlights the relatively weak security value of existing knowledge-based authentication approaches. As these industries increase the security of their services, fraudsters will inevitably use the same readily available information to exploit process weaknesses in other industries. In the face of widespread public dissatisfaction, governments and regulators are similarly increasing the scrutiny and punishment of organizations that fail to implement appropriate protection of customer information and assets.

Making the Case

In some cases, it is also inevitable that skepticism and uncertainty can unduly influence decision makers. It's important to note that the impact of voice biometrics on an organization's voice channel is not confined to any one area—it provides tremendous cost savings and revenue growth opportunities. The cost savings opportunity for voice biometrics is split between direct customer service operational costs and the wider costs of compensation and fraud. Customer services teams will need to decide whether the capacity created is realized as headcount reduction or repurposed through various measures, including the following:

  • Agent handle time (AHT) reductions: A significant portion of agent call handle time today is spent in the welcome, identification, and authentication processes. With voice biometric authentication, call time decreases because the requirement for customers to remember information is dramatically reduced.
  • Call volume reduction: Using voice biometrics, the requirement for customers to remember information is dramatically reduced, significantly increasing first-time success with an attendant, and, thus, a reduction in call volume.
  • Increased self-service usage: Voice biometrics leads to increased self-service usage because an increase in the automated authentication rate will likely lead to a similar proportion of callers using self-service features when available.
  • Complaint avoidance: A significant reduction in complaints related to the authentication process occurs because customers are more likely to pass the authentication process instead of being denied service for failing to pass a cumbersome artificial test.
  • Loss avoidance: With more customers enrolled in voice biometrics services, there is less opportunity for imposters and a smaller proportion of calls on which other detection and investigation methods can be concentrated.
  • Employee retention attendance: An often-overlooked issue is the stress and anxiety experienced by agents using existing authentication methods. With voice biometrics, the proportion of calls requiring manual authentication significantly decreases It is not uncommon to see 5 percent to 10 percent reductions in attribution and absence in call centers between the start and maturity of a voice biometrics solution.

Moving to Implementation

But with voice biometrics, there are some important considerations for approaching the design and implementation that are worth considering. However, companies must not treat it as a black box. The solution will only be as effective as the people using it and the process on which it is dependent.

When asking for customer consent, word choice and positioning in questions has the biggest impact on the success of the program. Involving customers and front-line colleagues early in this process is essential for success. If agents are driving the conversation with customers, they must believe that the solution works so the process is easier for themselves and their customers. Publicizing the results so stakeholders know the reduction in manual authentications following enrollment efforts can build infectious advocacy in your organization and call centers.

Voice biometrics is one of the key enabling technologies of the next generation of customer contact experiences. By delivering security that doesn't create an obstacle for customers and support agents, it enables significantly richer human-to-human interactions where secure authentication occurs in the automated self-service solution, and agents focus on meeting callers' needs rather than complying with processes. In doing so, voice biometrics helps strengthen and deepen the relationship between consumers and companies.


Mary McKenna is director of product management at Interactions. She has more than 20 years of experience in product development and delivery in search, big data analytics, and customer service automation markets. She has created, managed, and motivated teams that build products using artificial intelligence technologies for industry and government and provided business strategy, solution design, product management, and key sales and marketing support.


Related Articles

Back in 2015, one in every 2,000 calls into a call center was fraudulent, but in 2016, that rate grew to one in every 937 calls, according to telecommunications security and research company Pindrop Lab's 2017 Call Center Fraud Report. The change represents a 113 percent increase in fraud year over year, according to Pindrop, and every fraudulent call represents a loss of roughly $0.58 for the business.

Posted May 12, 2017