In 2020, Gartner predicted that by 2023, 30% of organizations would use at least one form of password-less authentication, thereby eliminating static, stored passwords— creating a market opportunity for biometrics to supersede passwords as we know them.
Callan Schebella, executive vice president, product management Five9, recently briefed No Jitter on the impact biometrics data will have on the future of contact center work as businesses continue to embrace digital workforces. In this Q&A, he explains why biometrics data makes for frictionless, more convenient customer interactions; how biometrics can change the way organizations approach safety; and how enterprises can use APIs to add an extra layer of authentication to their legacy technology.
Responses have been edited for conciseness and clarity.
How does Five9’s latest version of its AI-powered Agent Assist capability improve agent performance and deliver a consistent customer experience?
CS: You [may] be familiar with the idea of guidance cards popping up in response to an intent within the conversation between an agent and a caller. The idea here is that you can make much more complex guidance cards. [For example] when you engage in a complex interaction, like a multi-step or high-compliance process—instead of the guidance card popping [up], the checklist pops up on the agent desktop and has a number of checkboxes. What [the checklist] is doing is listening to the conversation and marking the boxes off. You can have a positive confirmation mark, a negative confirmation mark. Then there are boxes you should and shouldn’t fill.
The idea here is [the agent] can work their way through that checklist and revisit it at the end of the conversation to make sure they get everything. That’s the main noticeable thing people will see with the newer agent assist capability. In high compliance environments, it’s not so much telling the agent what to say but making sure they’ve completed all the steps necessary to have an authorized or successful transaction.
The worst thing that can happen in that environment is using a traditional quality management approach and detecting later that an agent has done something wrong all day. Now you’ve got to go back and remedy that—which can be problematic. The idea with the checklist is in the analogy—you can invest your money in the panel shop or make better drivers.
As enterprises continue to embrace digital workforces, how will biometrics data have an impact on the future of contact center work? What modern methods can enterprises improve upon to increase their level of protection?
CS: A substitute use case for biometric data in the contact center is a replacement for identification verification (ID&V) processes. Most contact centers have an ID&V process, ranging from call biometrics to out-of-wallet questions, pins, and things like this—all of which are tedious and perhaps not as secure as a biometric that travels around with you. The way biometrics impacts agent work is, it’s changing the way a contact center looks at a typical hour of agent time. It’s changing what that hour is being spent on—in the same way that automation in general is changing what agents are doing.
Biometrics is an enhancement of automation, in terms of reducing ID&V. [For example] you may reduce your average handling time by 30 seconds, because you don’t have to ask all those [security] questions. I’m reducing handle time by getting rid of the identity component. In non-call environments, such as webchat, live chat, or SMS-based interactions, where you need to escalate the call to a biometric in order to continue, you’re starting to mix the modality.
Why does using biometric data result in smoother, more convenient customer interactions? What benefits does it provide to the customer?
CS: Buyers of automation systems often have a biased view of how their customers think about them. For example, [the business may say to a customer] ‘tell us your member number,’ then the customer says, ‘I’ve got no idea what my member number is, this is one of the seven hotels I stay at—or one of the four airlines I fly.’ Then [the customer] is rummaging around, looking for the card, and trying to find their membership number. Then the business says, ‘I don’t understand—we send these cards out and assume everyone keeps them.’ Like no, that’s not what happens. [Customers] lose cards 10 minutes after you send them—biometrics is something a customer cannot lose.
How can the use of biometrics alter the way organizations approach safety?
CS: Here’s an example of the tele-clock scenario. Five9 has a big user in the APAC region that uses biometric tele-clocking in the contract-to-hire labor management business. One of the challenges these companies have, is they don’t even know who employees are, and at that point, you’ve already left their organization before the paperwork gets processed. If the business uses a biometric system, the biometric becomes an employee’s source of truth.
Some companies get caught up in issues related to security and the safety of their people because sometimes an employee may not be who they say they are. Let’s say I have the job to clean an office building at 10 p.m., but I can’t make it, so I send my cousin. By using a biometric, the employee has to get to the building, pick up the phone, log on with their voice, and log off when they leave. So you know who came in, you know it was them, and your duty of care is there.
What are some ways enterprises can use APIs to add an extra layer of authentication to their legacy technology?
CS: One of the challenges legacy technology platforms have is overlaying biometrics. Ultimately, you need to gain access into the audio channel to be listening, to be able to implement biometrics. At Five9, we make our platform available so developers can add a biometric engine in a seamless way. We do that through our VoiceStream API, which allows a third party engine provider to anywhere on the public cloud to go through a process for getting authorized to run on the Five9 network. Then the customer running the program can go in and essentially give permission to any of these approved vendors to talk to what’s called VoiceStream Access. So it allows them to listen to the conversation happening between agents and callers with just one click. Also, getting access to that audio data is very valuable, and you have to do it very securely.
What is one of the challenges you’re most excited to keep addressing at Five9? Why put your time and energy here?
CS: It’s not specific to biometrics. But, in general, if we look at what’s happened over the last few years in the contact center industry, you see this real emergence of accessible artificial intelligence or practical AI. That means a business could implement AI without having to spend seven figures. Now we’re talking about implementing things like natural language processing models at a tiny fraction of the cost.
Things are moving so quickly—and one of the big challenges left is, how do you build a natural language model unless you are skilled in doing that? [At Five9] we’re focusing on making model development and management accessible to anybody through no-code interfaces, drag and drop interfaces, and wizard-style interfaces in the browser. I think if we can solve for that, we’ll be able to see a much more widely adopted use of AI across the contact center industry.
Five9 was named a finalist for the Best in Enterprise Connect award for the latest version of its AI-powered Agent Assist capability, powered by Google CCAI, which provides call transcripts and summaries. Why do you it was recognized as a solution for contact center challenges?
CS: With the latest version of Agent Assist, Five9 continues to move forward on its mission to democratize AI and make it practical to use in contact centers. We are focused on delivering solutions that make AI persistent throughout the customer and agent experience and tools that make it easier for contact center teams to build, deploy, train and derive insights from the AI.