Authenticating People and Identifying Fraudsters with Nuance Communications’ Simon Marchand

Hello and welcome to the ID Talk podcast. My name is Peter Counter and I am the editor-in-chief of FIME Biometrics, the internet's leading source for biometrics news, interviews, and thought leadership. On this episode of ID Talk, I am pleased to present a conversation between FIME Biometrics' own Doug O'Gordon and Simon Marchand, Chief Fraud Prevention Officer and Director of Product Management for Nuance Communications' Security and Biometrics Division. They start on the topic of biometrics in healthcare before digging into Marchand's area of expertise, authenticating people and identifying fraudsters with voice biometric software. They touch on the high uptank Nuance is seeing from banks and how call center as a service models have affected security. Nuance is a leader in voice and behavioral biometrics, and Marchand gives us a great tour of how those modalities play off of each other in order to create seamless customer experiences that are protected by powerful and passive fraud prevention. It's an informative interview about voice biometrics across a wide range of markets. So without further ado, here is Fyme Biometrics' Doug O'Gordon in conversation with Nuance Communications' Simon Marchand. Right here on ID Talk.

Hi, Fine Biometrics listeners. Doug O'Gordon here. I'm with Simon Marchand. He is the Chief Fraud Prevention Officer for Nuance Communications. Hey, Simon.

Hey, Doug. Thanks for having me.

Yeah. Hey, well, so in full transparency, I met Nuance at HIMSS. I met all the marketing people. We had a great conversation. I went through your demo and saw how the doctor records everything on, what's that called, DAX?

Yeah, yeah, DAX, exactly.

But I did not meet you, so I'm meeting you for the first time here.

And yeah, we obviously didn't have a chance to meet them. Nuances is known in great part because of its work in healthcare, but I'm on the other side of the business, working in security and biometrics. So using kind of the same core aspects of our technology of voice recognition and transcription, but applied specifically to authenticating people and identifying fraudsters. So same tech, same big business, but, you know, applied for security and fraud prevention.

Yeah, and I didn't know if you knew this, but we cover all biometric modalities across all vertical markets. So, you know, definitely, I, you know, my background is security, um, and, and I, you know, it's great to talk to you because that's really the sweet spot, I think, of biometrics and identity is making sure, you know, the bad guys don't get through. Well, let's talk about one identity question, a fraud identity question in the healthcare market. Currently, what's the biggest identity-related challenge facing the healthcare industry right now? Now with this big digital transformation occurring.

Yeah, I think when we're talking about identity, regardless of if it's healthcare or others, I think the biggest trend that we see right now is the whole shift to digital identities. During the pandemic, we have seen how challenging and complex it is to, you know, do business remotely because most organizations were not ready to do this. They were not ready to, you know, to have this patient engagement remotely, for example, or to have like a bank customer kind of relationship fully remote. So when you're talking about identity, it's really how can I make sure that you're the person you say you are, while never having the opportunity to necessarily see your IDs or have you in person in front of me, and while not having control over what kind of device or hardware you have available to you. So we see a lot of governments, a lot of states, you know, looking at digitizing their identities, making sure that you move from that little plastic driver's license or insurance card or whatever you want to have to something that you can carry with you on a phone, but that will also be used without needing to have that phone with you. And that's where biometrics comes into play. If you can enrich the basic aspects and attributes of your identity with some biometrics factors, then you can lock your identity behind it, but you can also access it from anywhere. And that's, that's really the power of using voice, for example, because it doesn't matter if I have the latest smartphone an old copper line, if I'm doing it in person, or if I'm on a Skype session or a Teams session, it will always work because it's all about the way you sound and not the device you have in your hand. So that's what we see on the identity side, that kind of shift and that enrichment of identities with biometric factors.

Sure, well, and then we're done with HIMSS now. HIMSS was a great show, but let's just keep discussing fraud. And where do you see the biggest opportunities for like a Nuance and using your technology across all of these markets that we could be talking about? What's your first market go-to, the one that seems like easiest solution for you?

Yeah, Doug, there's so many markets we can play in. I guess Very naturally, banks come to us the first because they have some obligations, you know, to comply with some rules and regulations depending on where they are, but they have that reflex to strengthen their authentication process because they have those steps that they have to follow, you know, proper KYC procedures, so know your customer, but you know, overall, they're protecting your assets, your savings. They need to be credible, so they cannot rely on old dated methods. So KBA, you know, knowledge-based authentication, where I ask you your name, your date of birth, your mother's brother's maiden name. All of that is useless. It's meaningless. That information is out there. Fraudsters can purchase it. They can obtain it. They can steal your mail, go on social media. They can do a transaction on the dark web marketplace. So really, banks, I think, are the first ones to adopt that kind of technology because you remove the need to ask questions to your customers. You just let them talk. Hey, I'm Simon. I'm calling today because I need to change my address. Good. We know it's you. It sounds like you There's a biometrics algorithm that's, you know, an AI-powered algorithm that certifies it's you. Now after that, the next step after banks is anyone ever that has to authenticate you. If you call your carrier, if you call your utility company, if you call, you know, the IRS, if you call an insurance company, all of these organizations need to make sure that they're transacting with the right person, and we see demand from all of these organizations. Basically, anytime you have to answer any kind of security question, you can imagine that there's a need for biometrics there, and it's the kind of organizations that we're working with, big or small. That's, that's kind of the benefit of having a cloud-first solution. You don't need to be a huge player with tens of thousands of agents in a call center. If you have 12, 15 agents taking phone calls, we'll find a way to make biometrics work for you. But yeah, pretty much anyone can benefit from it.

And explain to the audience how, you know, how the hardware is not as necessary, right, on this. Like, the capture of this on variety of hardware is so much easier.

So it's depending on where you deploy it, I think, is what's going to drive that, you know, the answer to that question. We see that the market is shifting more and more to contact center as a service, you know, so CCaaS platforms that are completely in the cloud. So you don't have those, you know, telephony boxes on-premise that we need to pull the audio from. Now we just do cloud-to-cloud integration. You want to do biometrics, you have your, your CCaaS platform already, we'll just enable Nuance Voice biometrics. Our clouds will be talking to each other, and you know, right away you get the benefits from it. If you have one of those old systems, which is the case of a lot of organizations, then You know, that's the benefit of having done this for 20 years now. We have the experience with pretty much any kind of hardware you can imagine, but it's basically pulling the audio that's transiting through your call center and just forking a copy to Gatekeeper, which is our biometrics platform. By doing this, we can listen passively to a conversation and do the biometrics authentication during the call. So you're just, you're talking to an agent and right away we listen to it and we can say, yes, it's Doug. You can issue a new credit card, change an address, give a new phone. But yeah, so it's fairly straightforward of a process. Now, others will decide to have it embedded directly on the phone. So you can open your app and say, my voice confirms my identity, for example, unlock some high-risk transactions. So you could use the hardware you have in your hands. You can do some web portal integration where you use the microphone from your tablet or from your computer. Or you could go in person, you know, you could be in a, like a financial advisor's office and you could have a little microphone, you could do it there too. So really there's a multiplicity of ways to capture the audio. I think what's important to keep in mind is as long as you have the proper expertise with you building that solution, there will be a way to get the audio and to make sure that your customers can benefit from it.

Where is this audio? Like, where is it? Where is it stored? How does it— how does that process work?

So it's all in the cloud. We focus on software as a service because it, you know, from a security standpoint, it's always more interesting to be able to push new algorithms to make sure that, you know, if a fraudster starts using deepfakes and deep voices, they can't breach through our systems. So everything is in an Azure cloud instance. Obviously, you know, Nuance being part of Microsoft now, it only makes sense that our systems would be in Azure., but we, we have everything processed in the cloud, so no hardware that you need to, to purchase unless you have a very particular situation. But yeah, otherwise it's all dematerialized.

Fascinating. I, I find this— are they using other multi-factors, other forms of biometrics on top of your biometrics?

So the biometrics we have is actually threefold, right? We don't have just voice. So voice is the sound of your voice. Can speak English, French, made-up languages, it doesn't matter. You will still sound unique, but we pair it with conversational biometrics, which is only the words you're using. And when you use the words, what's really interesting is not only can you certify who a person is, but you can start detecting if someone's following a script, if they're being coached through a process, if a fraudster is asking for something in a very odd or suspicious way., but not just if they're asking for it, if they're typing for it, we can start looking at how you interact in a chat session. We could, you know, virtually look at your emails, for example, to determine if it's really you because punctuation, emojis, spelling mistakes, all of this can be used to help profile someone. And then if you're interacting on the web, well, we can use behavioral information. So if you're using a phone, it's how you're holding that phone, how you're swiping on the screen, how you're typing with your thumbs. If you're using a mouse, well, the way you move your mouse, the way you click on buttons, All of that is part of the behavioral print. So you can already see if you use voice and words and behavior, virtually we're covering all of the channels you might want to interact on. So we really focus on these 3 technologies and of course around it, you know, a huge portfolio of smaller detectors that we call them. So making sure we can detect if a recording is being used, if a synthetic voice is being used, if there's a bot involved in the transaction. So all of this is also factored in.

And that's all done behind the scenes?

All behind the scenes.

As I'm talking to you, if I'm a bad guy?

Yeah, exactly. And if you're a bad guy and we have your voice, well, your voice you're stuck with for the rest of your life. So if we identify you once, you can't come back. We'll detect it, and we're talking seconds in that case, right? For a fraudster, 5-6 seconds is more than enough for us to say this is a suspicious individual., and then let the agent know that, you know, they might have to escalate to professionals.

I always look back on old movies and go, that would never happen back then. Like, there's a Catch Me If You Can. Frank Abagnale never would have been able to do everything that he did. Yeah, you know, I actually can't do it anymore.

Exactly. And I've met with fraudsters, you know, very prolific fraudsters that, you know, stopped doing it, but that threw everything they had at me. Trying to find how they would break through a biometric system. And hey, what if I did this? What if I did that? What if I brought this kind of technology? And to this day, I'm yet to find someone that has the right way to completely circumvent our security measures. So I'm always very proud of that.

So what's your next trade show? Where, where are you headed to? Like, what's left on the plan for Nuance through 2022?

Nuance will be everywhere in 2022. It never stops. Fraud is a, is a huge, huge world, but a very, you know, tightly knit community. So you have to be in those communities to, to really bring your message. So, you know, I've done some incredible work with the Association of Certified Fraud Examiners, which I'm pursuing this year. I'll continue to work with, you know, them at the global scale, but also the more local and regional scale. So, so still a lot of work to do on that front. But if you hear of a fraud association somewhere, there's a good chance I'm going to be there. So we might bump into each other if you're going to be in a security or fraud conference in the coming months.

Yeah. And then, uh, what's the best way if anybody who's listening wants to get a hold of you or find out more about Nuance? Uh, what's the best way to, to do that?

Always happy to have people reach out on LinkedIn. So They just have to look me up, Simon Marchand, M-A-R-C-H-A-N-D. But you can also go on nuance.com/fraud if you know you want to have something a bit more formal or get more information. So, you know, whether it's white papers, recordings of webinars, or conferences that we have given to speak on that topic in more details, nuance.com/fraud would be the destination for that.

Great. Well, I thank you for your time and, uh, have a great rest of your week and hope to run into you at a fraud show in the near future.

Oh, I'm sure we'll see each other. Thanks a lot for having me today, Doug. Appreciate it.

Thank you, Simon. Cheers. Bye-bye.

And so concludes our interview with Simon Marchand, Chief Fraud Prevention Officer and Director of Product Management for Nuance Communications' Security and Biometrics Learn more about the topics discussed in this episode by visiting Nuance.com. And for everything else biometrics and identity, stay posted to FindBiometrics.com. I would like to thank Simon Marchand for joining us on today's episode, and thank you to Doug O'Gordon for conducting the interview. Our podcast theme music is by Logum Rad. I have been your host, Peter Counter. Thank you for listening to the ID Talk Podcast.