Cerence Inc. (CRNC) Earnings Call Transcript & Summary

On a lighter note, Mark, I'm engineer, Mark Gallenberger, our CFO, is an engineer. Leanne, who's our General Counsel. He's an engineer. So we have a lawyer, who is an engineer. My head of IR, Rich, he's an engineer. So we are all -- a bunch of engineers really passionate about technology. We really are, okay? And that's what we do. I have -- from the day I came in as a CEO, I've started companies here in Silicon Valley. Over the last 25 years, I led other private companies. And the simple mantra that I follow is innovate or die. It's as simple as that. And our priorities are very focused, number one, product and innovation. Every single morning when I wake up, I think about that for our company. We have a strong CTO. We have a strong product management team, basically also thinking, right, and working with our customers. Our second priority is speed of execution, which is executing across all functions and domains and so on. And our third focus is cost. So these are the 3 simple things that we focus on in the company every single day.

No. That makes a lot of sense, and it's very helpful. I wanted to get into one of the debates that's been ongoing in the investment community about the competitive landscape and also opportunity for Cerence going forward, and what roll is big tech going to have on the auto industry? And how would that potentially impact Cerence? And Cerence's articulated a view that there's going to be coexistence between your technology and things like what Amazon and Google and Apple can potentially offer with Siri and Alexa and those sorts of services. Can you be more specific and talk in more depth about how you see Cerence coexisting with some of these other big tech voice technologies and connected car offerings?

Sure. So I'll start, and maybe mark, you can jump in as well. So the -- firstly, voice, AI, conversational AI plays is a very important role in the car. We already talked about that. We are a specialized platform for the vertical application, which is the car, right? Many big tech are more horizontal basically because that technology exists on the phone, in a TV, on a set-top box remote, in a spot speaker, et cetera, et cetera, in the car. The -- from an OEM standpoint, it has -- there has been a consistent message that they want to own the user experience. They understand that the car is getting more and more digital. And owning that digital experience is extremely important to the OEM. They don't want to outsource the digital experience to somebody else because if you outsource that digital experience, you are basically giving up -- you're outsourcing the soul of a car. Because if you believe that the car is getting more digital, that the digital experience is a soul. So #1 reason, they want to own the digital experience. Second thing is that through that AI, conversational AI and other elements of AI, that technology enables the OEM to understand the -- what their drivers and the consumers are doing in the car. What they -- what functions they are using, what transactions they're doing. And with that, basically, there is an opportunity to improve the product, there's an opportunity to monetize further areas. If you outsource that whole piece to somebody else, you basically kind of lose that control. Cerence's model is OEM-friendly, meaning our model is B2B, and we basically provide a white label opportunity, white label product for the OEMs. Having said that, the coexistence is absolutely necessary. I am not saying that this is -- it's something that should not be talked about, et cetera, it's one or the other. Absolutely not the case because a consumer spends 22 hours outside the car and maybe 1 to 2 hours inside the car. And the consumer wants a seamless integration of his or her digital life outside the car to the digital life inside the car, right? And so what we are trying to basically do is try to create that bridge, try to give OEMs what they want from a private white label AI platform, conversational AI platform standpoint. Full UI control, full access into the data, full ownership into the data, et cetera, et cetera, and monetization of that data. But also a bridge to the multiple big tech ecosystems that should coexist in the car. And because if I look at my digital life, I use Apple devices and services every day. I use Google devices and services every day. I use Amazon devices and services every day. I use Microsoft Calendar, Outlook, et cetera. I don't want one or the other. I want all 4. If I am a consumer sitting in China, I don't care about Amazon or Google, I care about Baidu or Alibaba or Tencent. So I want to bridge from my car to Alibaba, Baidu, Tencent. If I'm a consumer sitting in Russia, I want to bridge to Yandex, not to Google or Amazon, right? So what we are trying to basically do is create this platform that enables what OEMs want, but also what consumers want, right? And I think it's very visible. As you know, a few months back, Google who had closed their API for Android automotive, have opened up the API to coexist -- coexistence of multiple assistant alongside Google Assistant, for example. They will not be doing that if this is not what people want.

Yes, that's really helpful. So Cerence helps the OEMs have that digital experience and own that, better control of the data. But the other thing, I mean, you were talking about as a bridge. But in our conversations in the past, you also talked about cognitive arbitration, where the first thing that the car does when someone speaks is clean up that signal. And then if necessary, route it off to a Google search or something like that. But Cerence technology is that front-end step and can provide the logic about what to do with the voice command, even if it ultimately does go off to one of these big tech companies?

That's right. So there is a cognitive arbitration that's built in that part of our product. And there is multiple layers of cognitive arbitration that we do. The most basic is wake-up word based arbitration. So you say things like hey Mercedes or hey Alexa or hey Google. And basically, it switches based on just a wake-up word. That's very basic. We also do intent-based arbitration, which basically means that the -- it's like an intelligent voice router, right, which basically routes based on intelligence. So based on what the intent of the conversation is, tell me about weather in San Diego. And so it will just kind of say, "okay, for whether, my partner is this." And so the business logic for that voice intelligent router will be controlled by the OEM in this case.

That's helpful. We talked about this on the earnings call, Sanjay and Mark, but I think it's really important. So I want to bring it up again and perhaps you -- so show more examples you can share with us compared to Monday on the earnings call. But I was asking about what kind of pricing and content per vehicle is Cerence realizing when it's the only voice technology in a car compared to when there is this coexistence with big tech. And the concern being, from some investors, that if Cerence has to coexists with these other big tech companies, okay, maybe you do coexist, but you're going to have to offer it at a much lower price because there's other companies there. And I was asking you, what kind of dollar opportunity per car are you getting in those instances where there's a coexistence compared to when it's just Cerence? Can you talk a little bit about some of those examples? Are you seeing similar economics when you're coexisting compared to being the only provider?

Yes. So we have not seen any major shift in the economics. We have co existed with Amazon on a number of platforms. We're in discussions with a couple of OEMs on the coexistence with Google, et cetera. And no major shift in economics, Mark, that I can share at this stage, right? And the reason for that is that if you look at the big tech business model versus our business model, it's different. Our business model is pure software licensing for the OEM, right, and providing them this white label platform, right, that they own. They own the experience, they own the data, all that stuff. The big tech, if I look at Amazon, for example, or others, their business model is -- the reason for Alexa to exist from Amazon standpoint is a very different business model, right? They're not trying to license Alexa. As a platform, they're trying to monetize the transactions that Alexa grabs when it's existing on a platform, right? So it's a different business model. And the coexistence kind of doesn't erode either side, right? It's kind of -- it's one of the happy way of kind of staying together and solving the big problem. Mark, I don't know if -- the economics part, I think, my answer is accurate. We have not seen any major kind of thing...

Yes. I think that's right, Sanjay. I think also, given some of the changes with opening up the APIs with Google, that's more testament to the fact that the OEMs as well as the end consumers don't want to be segmented to just a single large tech solution. And so that coexistence is evidenced in that. And as a result, we do believe that there's going to be the ability for us to continue to have that -- enjoy the same economics that we've had historically, even with the coexistence story.

That's helpful. I do want to make sure we speak to some of the other end market opportunities that Cerence is pursuing when we think about transportation more holistically. And there are some instances where that's outside of the cars and elevators is something that the company spoke to on the last earnings call. You said you're in proof-of-concept demonstrations at 2 of the 5 largest elevator manufacturers globally. Elevator is not a market that a lot of tech investors are familiar with. They're certainly not so that I'm all that familiar with. So can you talk a little bit more about what kind of design in cycle and time to market there could be in elevators? And any sort of commentary about the potential revenue opportunity from elevators for Cerence?

Sure. So frankly speaking, Mark, even I was not thinking of elevators. This came because of COVID-19, believe it or not. So when COVID-19 hit in March, I used to go to my quarantined office every day. And our offices in San Mateo are on 6th floor, and I would not want to touch the button, both inside in the hall and in the elevator. And so I would use my key or my access card or, et cetera, et cetera. And I was talking -- brainstorming with my product management team and people were having similar kind of experiences. So we started kind of studying the elevator market and started to kind of talk to some of the elevator companies. And it was very interesting kind of nice cross-section that they were also thinking the same thing that we were thinking. So we quickly kind of created a product around it. The product is different. So it's not just -- people may think that it's just the same product as a car, voice assistant, simplified, that's not the case because the use cases are very different. Let me explain a couple of them, right? So one is you always assume that all the people sitting in the car speak the same language. So although we support 70 languages, in a car, you get -- you enable English U.K. or Australian U.K. or -- sorry, Australian English or Chinese or Mandarine or whatever, right? And -- but in an elevator, the product should be able to support multiple languages at the same time because there are -- there could be multiple occupants in the elevator speaking different languages at the same time. That's point #1. Point #2, voice biometrics plays a very important role because today, if I go -- I use my access card to get access to my floor. And I don't know about you, Mark, but I have absolutely left that card and have been stranded at an odd hour in my lobby saying, okay, what do I do? Why can't I use my voice bio as a way of authenticating me to take me to my floor? Third, why do I need to know the POIs? Why do I need to know which floor Cerence is located at? Why can't the elevator figure it out and all I say is, "hey, take me to Cerence", right? So there are all these use cases that are really unique and interesting as we dig deeper. We work on our product for about 4, 5 months in a very fast way, prototyped it. In August, last year, we started talking to elevator companies. We picked the top 5 elevator companies to start discussing with them. We successfully finished our first POC in December with one of them. And we are in the POC with the second one right now and then in discussions with the others. And we expect that the product will go into -- fingers crossed, we've been at it obviously, we have not won it yet. So I should not make too far-looking statements, in second half this year. So they'll start offering that as a product. The exact scope of the opportunity, we're still assessing it, Mark. It's a little early for us. There is about 2 million elevators shipped every year. There is a very large installed base of existing elevators. Our product can be retrofitted into existing elevator base as well. So we're not going after only the new elevators that get installed every year. And again, just like cars, you have low end, mid-end, high end. Same thing with the elevators, you have the low, which is the residential, 0 to 3 floors. So we've learned a lot about the elevator market in the last 1 year.

Yes. And do you think about the dollar for elevator opportunity? Is that something similar to what you may be doing at cars of mid-single digits per unit? Or is it completely different for elevators?

It's completely different. It's completely different, Mark, because there will be a -- if you look at our announcements that we made at Cerence In Motion, there will be a hardware device and software. So not all elevators are enabled to run AI, right? So the economic opportunities is different. And remember, like a car where you think of 10-year life, elevator life is double to triple of that, basically. So there is a SaaS monetization part, right, that we are trying to basically bundle into our product offering.

Okay. Well, sounds really interesting, and I think investors and myself are all looking forward to seeing how that develops. We've gotten to a few of the questions already that the audience had sent in so far. But I did want to try and get to one more if we could, before we run out of time. Question is about the Mercedes plans to work with the NVIDIA DRIVE platform in the future. Is that something that Cerence would be a part of? Do you participate on that platform? Or is that a potential risk if you're not involved on that development plan?

We're absolutely fully integrated and working together with NVIDIA and Mercedes. And as we announced earlier last year, earlier or mid last year, we were -- we won the next-gen of Mercedes platform as well. So yes, we are.

Let me just slip one last question we got from the audience in. I know we're out of time, but I think it's a quick answer because I know Sanjay, you and I have talked about this before, but the voice authentication for secure payment, how secure is that? Could somebody fool that with just a recorded voice?

It's very secure. I mean it's in use today for banks. When you call a bank and bank kind of recognizes you with your voice. So it is in use at the bank level today. Nuance does provide that platform, right? So we don't deal with banks directly, but our parent company did. And so it's fairly secure. It's not at the same security level as face ID or a finger ID, but it's getting there, right? But in cases where you need that level of authentication, you do 2 steps. So meaning, you take the WIN number with Sanjay's voice to conclusively say this is Sanjay, for example. So what you do is you do 2 steps, right? You add a factor, you do -- you add one more factor to the authentication.

Understood. I'd love to keep discussing everything, a lot of interesting things to come, but we are unfortunately out of time. So Sanjay, Mark, really appreciate you both speaking with us today.

Thank you so much.

Thank you, Mark.

Operator, this concludes the session.