ElevenLabs is obviously extremely well known for voice AI, for text-to-speech. For I think that beautiful intro we just got was actually on ElevenLabs, a little co-branding there. Um, and I'm wondering, you know, in the last discussion we heard this, uh, you know, topic of is the text box the best interface for AI? And I would imagine you have a take on how no, you know, voice is the best interface for AI or voice is the best interface for computing going forward. I'm interested, like, what do you think are the best use cases for voice AI and where do you see it, you know, today, a year from now, five years from now and beyond.
First of all, thanks for having me here. Great to see you all. And I actually didn't know this was voice generated, but it had a great pronunciation of my surname. It's actually hard. So I'm happy.
You guys train on your last name specifically.
It's, we should, I don't know if we do. So we'll, we'll definitely do now going forward. Uh, but the, so as a company, one of the key things we, we are aiming to solve is how humans and technology interact. How you create with technology and make it seamless. How you interact with technology and make it seamless. Um, and in general, to your, to your question, we think voice will be one of the key interfaces for interacting with the technology across. Um, from the simple pieces like interacting with the, uh, personal agent to help you go through the day where it can be on your headphone and be able to guide you, okay, through to education. That's one of the ones that I'm probably the most excited about, where in the future the combination of what LLMs allow you and what voice will allow you is that you'll be truly immersed with learning the given experience where you'll be able to effectively have, have your personal tutor on their phone helping, helping you across. Um, then the third one is of course for voice and for the language barrier to break, we need to figure out how to be able to speak across different languages while carrying the same intonation, emotion, voices, um, which, which, which, which will be a big, a big shift. And then in general, how we interact with everything around us, whether it's the, uh, the, the, the laptop, the phone, the robot in the future. And I think robot maybe is the easiest example. Of course, this will be voice driven. There's no other interface. And, uh, you today, maybe it's a year or decade, as Karpathy said, of, of agents. Of course, there's on the horizon the decade of, of robots. And, um, and I think here too, the most common interface will be, but it's not, it's not going to be only interface though. So you have...
Yeah, I mean, it's interesting you brought up those use cases of like a personal assistant, a tutor, and I guess a robot, you know, uh, house, house helper or Droid or something like that. Like, is your mental model that basically anything that today is something where you could have a human counterpart, right? A human tutor, a human assistant, a, you know, human, um, in your house, like you're going to fall to that voice interface as the most natural way to do it because we as humans are already used to using voice for those things? Or are there things where, uh, today, you know, voice isn't used at all really, but it's something that we're going to expand into going forward?
Yeah. So, so first of all, for sure, I mean, are we already seeing that? And I think that's the easiest one and the most immediate one is how customer experience, customer support is just changed and elevated where you instead of calling and trying to rebook your, your ticket and going for this IVR flow of, uh, click one, click five to get to the steps and waiting for the number of minutes. Um, you, you will have an agent that fully understands you, can guide you to the response and, and go through.
I, I wanted to get into that actually because we talked a little bit about agents on the phone and the, the, you know, calling United Airlines or American Express or something like, what percentage of customer support calls today are actually, you know, managed by a voice AI system or an agentic system or whatever you want to call it versus, you know, like, you know, IVR touch buttons and, and, and how do you see that progressing over time? Like, is that exponential curve going like this every year?
Yeah. Yeah. I think it's, I, I think the exponential curve is, is going like this. Especially like this year, we've seen incredible adoption where it's like Cisco, Twilio, Deutsche Telekom, all of those kind of leaning in quickly to rebuild how you interact with, uh, with help of voice agents. Yeah. Um, and I think the, you're, you're right, the IVR flows is still a big part.
Can I call today and get a voice AI agent on the phone?
You can, uh, you can, uh, from, we, we did our little summit yesterday, uh, as well and, uh, uh, one of the great ones was voice ordering with Square. I can call Square and actually order food delivery through, uh, help of, uh, a lot of the, the shops that work around with Square and actually do it through, through voice. I actually recently, there's a, if any of you are from London or traveled to London, there's an amazing restaurant called Zephyr. It's a Greek restaurant. Okay. Um, uh, where you, we, we, we worked with, with the company supporting that where to my happy moment, uh, I noticed that on their website they actually had ElevenLabs agent that you could call and actually book, book a spot there too.
So you can book a reservation at this restaurant in London with a voice AI agent.
And it connects, of course, to your calendar, your appointment scheduling, um, which is, which is great.
And when do you think we hit the tipping point where like the average customer service call goes through a voice AI agent? Like the median, the 50% point, whatever you want to call it.
I think over next 18 months.
Next 18 months. Yes. Okay.
So like mid-27, I call the average customer support is handled by voice AI.
Exactly. And I think, I mean, this is the most immediate one, the one where we see the highest ROI and value. Some of the other use cases we see as the kind of, the kind of where the, the future is headed. But to your point, there's, there's definitely one flavor of, of, of your, your point, which is how you can do, um, things more efficient through voice, uh, with the existing services. But there's also the second theme where you can do things that weren't possible ever before. Um, uh, uh, one, one of a good examples was our work with Epic Games where we brought effectively Darth Vader alive in Fortnite where millions of players could interact with Darth Vader live, uh, throughout the game, which of course is not possible in any other way.
James Earl Jones voice, right?
James Earl Jones and his estate worked with us and is such an iconic and incredible voice. And we think like in general that, that concept of like what was never possible before where you have incredible voices, talent, you can now shift them to be not only static but actually dynamic delivery, personalized and different for all the users. Something that you, you are already doing in many ways at, at Voli as well in an incredible way. Um, I think this will be a big...
I have to ask as someone in gaming, right? I mean, the Darth Vader did famously go slightly off the rails and maybe say some things he shouldn't have to various players online. Like, how, how involved are you guys in that? How much of that is something you're protecting against going forward? Like obviously with the IP partners and so on, they really want to protect, I guess, I guess you couldn't say it's the squeaky clean image of Darth Vader, but a certain persona of Darth Vader. Like, what's, what's your sort of go-forward plan as you license more of these IPs and voices and things like that?
Yeah. So, so in that project we are specifically involved on the, on the voice side. Yeah. Uh, but in general as you think about those deployments, uh, and, and that, that's the most, the most common theme is you don't only need the, the voice or the interactive experience. That's kind of one part of the equation. Then there's two other big pieces to really make them valuable. The second one is how you integrate that with other systems and actually bring the, the knowledge base, the data, the business logic inside of the system and how do you make it interact with the real world. And the third one, which is the one that you mentioned, is how do you now deploy that in production with the right testing flow, right evaluation flow and then monitor over time as we, as, as it behaves, put the right safeguards and evaluate and adjust that based on that case. Um, so that's something that we, we spend a lot of time on with, with a lot of players.
So testing more and evaluating more so Darth Vader doesn't go off the rails.
Yeah, it's any, any, and even, even like in, in a customer experience, you don't want it, for example, to shift and speak about politics. You want it to keep it on the, on the...
Even if you say ignore all previous instructions and...
Exactly, even if you, which is actually harder to say when you have like this, you probably see the prompts with a lot...
Hard to do prompt injection with voice.
And the Unicode characters say them out possible.
Okay. So maybe voice is slightly less susceptible to prompt injection than LLMs. I'm interested with like, it's a good segue into sort of celebrities and celebrity voices because I know you guys announced, I believe yesterday, you're setting up kind of a marketplace for celebrity voices and you have Michael Caine on there. And, uh, you know, at our company we build voice AI games, as you know, I would love to use Michael Caine in our game. Uh, uh, yeah, we can, you know, obviously it's a high gravitas, we can make a Batman game with him, something like that. Like, what is the process between, oh, I want to use an AI version of Michael Caine in my game to actually, you know, shipping and, and what, which parts do you guys take care of and which parts do I need to go off and deal with Michael Caine's people, I guess.
