Victor Riparbelli - Co-Founder & CEO at Synthesia.io

0:00Hello and welcome back to Scaling Europe. We are here live with Victor from Synthesia. Victor, it is a pleasure to have you. Thank you for joining me. Thanks for having me. How are you doing today? I'm very good. I'm very good. Shall we just jump right into it? You released Synthesia 3.0. When I was reading about it, I was watching your video about it. It seems truly transformational in the sense that you're trying to really disrupt video in an entirely new and different way. Can you touch a bit on some of the key features, some of the key highlights for those of people who may have missed it?

0:32Yeah, for sure. So when we started the company back in 2017, right, which is a while back now, there are basically two big ideas that we were interested in. The first one was that as we can create AI, create content with AI, not just video, but like speech and text and all the stuff we've seen today, the marginal price of producing content is going to go to zero, right? Not just in dollars, but in time required and in skill required. And the way we always thought about this is just like, you know, 10 years ago, you could write a book and publish a book from your bedroom. You don't need anything else in your laptop.

1:04That's going to happen for video as well. And that in itself is a big, big opportunity. I think that's kind of like, we're still early in that, but that's been most of the company up until now. Just making video faster, better, but still fundamentally the same type of video that we've always engaged with, right? The second idea was this idea that what always happens when we as humans invent new technologies is that we create new media formats that are native to those technologies. The question there is, what does video look like when video is not something you record in the physical world with a camera and a microphone, but something you program, something software creates?

1:38That creates a whole bunch of new, exciting possibilities. And I think just like the difference between a newspaper and a website is today pretty big. They're very different mediums. Like the first website looked like a newspaper on a screen because that's kind of what we could imagine, right? And with 3.0, we're taking our first real step into innovating on the video format itself. And fundamentally, what it means is that we're moving away from video being kind of a broadcast format, which means that, you know, you record it or you generate it once and then everybody watches exactly the same thing.

2:10So being something that's interactive, it's conversational, and it's a very different experience for everyone who interacts with the video. unlike the kind of traditional broadcast format. So in practice, what this means is that you can insert these video agents into your videos. So let's say you're creating a training video, which is a very common use case for us. You're training your sales team on a new product that you just launched. What a lot of our customers will do today is they will create lots of great video content. The sales team will sit down, they'll watch the videos and hopefully they'll understand the content they'll be ready to go out of and sell that product, right?

2:40Of course, we all know that in reality, that's not necessarily always how it works. And so what you could do with a video agent, you could say first you watch this like five ten minute clip and then by the end of it you go into a simulation with an agent that pretends to be a customer and then you know you do 25 turns with the agent and essentially you have to overcome objections and you're basically being tested in real time you've actually understood the content or if you've just like consumed it right which there's a very big difference on and that's going to be a big game changer because before this who would test that you actually know what you're been consuming That would probably be your manager.

3:14It'd be a coach, be like a real life session somewhere, or you'd just be thrown right into it with customers and hope that you're kind of well-prepared. So it's a very different paradigm from the traditional video. Now, with these agents, you could do lots of other things, not just about like role-playing. Some of the stuff we're excited about is building these kind of recruiting agents where there's a lot of demand, whereas you could go and create a video about a job, give people a case study where they're talking to it live on the call, which helps the recruiting team, gather more information upfront, making better hiring decisions.

3:47In product marketing, you can imagine that you're watching a video about a specific feature and you have the ability to ask questions to the video. So maybe the video doesn't cover all the questions that you have. Just like with a salesperson, you can ask a question. You can now ask the agent a question. So it's a pretty big shift from just broadcast video. The way we thought it out though is that this should be something that's additive to the existing creator experience. And so I think once we launched this, hopefully like in a month's time, it should be something that's very easy for people to pick up and use and essentially like the same type of person who uses some these here today should be able to to engage with this like very quickly outside of that there's also a whole bunch of other just improved into the core products like better avatar models better voices kind of workflow optimizations enterprise readiness but the agent piece and moving video from broadcast interactive is is sort of like the the main headline of what we announced last week yeah i think on that first piece moving from broadcast interactive you know you wrote on your newsletters that static video would be a reddit of the past which I thought was really interesting and for me like I'm a my job is essentially content creation now you know my background was in product and working in tech and building products but now I've just become like a social media manager and a content creator and for me I'm in I'm in the world of news and breaking news and stories and for me I've been thinking about how many ways that I could change my content using this and making video interactive and turning an article that you might read on a publication into a video and not just a video but an interactive video i think i think it's going to be truly game changing and i'm super excited to like try it myself and try and create content and do media differently using these tools um but i also wanted to talk about some of the agents you know there was a quote of yours and again in the newsletter where you talked about you know the real breakthrough is a new category of data video agents operate with specific knowledge of your business they can capture data in real time feed it back into your systems and take any actions you assign and then you talk about easily automating and scaling repetitive processes and this is for me reading that that was like okay okay, you have been talking about moving into the agentic world for a while, but this is like pure agentic play, even not necessarily related to video.

5:48But I could also, I could imagine this agent being plugged into CRM systems, right? Maybe it's a prospect is speaking to a video and then your agent is taking information and feeding it into a CRM on the back end. Are you imagining those kinds of use cases? For sure, for sure. I think the intelligence layer is obviously not something we're building ourselves, right? I think that's not what we're uniquely good at. What we're good at is the presentation layer, essentially. What we've built is a great editor. We've built great workflows around that, which goes further than just the creation experience.

6:20It's about from idea generation, creating a draft of the video to content management, translation, version, updating the content after you publish it. We have a publishing platform with our own video player. It's a collection of a lot of different things we've built. But I fundamentally, I just think that if you look at a lot of the, if you look at the AI landscape today, it has a bit of remnants of like when a company kind of ships their org shop, right? Either the modalities is kind of separate companies. You have like voice companies, you have video companies, you have, you know, transcription companies.

6:54Like all these like foundation models is like built by different companies, which makes sense because each of them are very hard to build individually, but it's not very natural to how we actually consume content, right? Like most of the time as a human sitting at the other end of all these models, you want to consume something that mixes all these models in different ways, right? I think we're starting to see all this stuff sort of converge now. But for us, I think there's a lot of these agendian use cases where you want a visual to guide you through it. And that's not just an avatar talking to it, to be very clear, right?

7:22That could be if you're learning about the pricing of a new product that your company is launching, you probably want to see a pricing chart explaining how it's priced for different customers, right? If you're learning how to, you know, clean the deep fryer in a fast food company you probably want to see a real video of someone cleaning a deep fryer or a really good air generator one right so for a lot of these use cases you want visual content there's also going to be a bunch of use cases where you don't need a visual context that if you're calling on your phone to order a table at a restaurant right you probably don't need like a full visual interface you can just say i need a table at 6 p.m but our bet is basically that there's going to be a lot of these use cases which are going to be driven by visual content and what we're building is essentially a presentation layer right where we avatar you can talk to, but it's also about fetching the right content at the right time.

8:08It's about generating content on the fly that's relevant to whatever you're talking about with the agent. And then, of course, building, as you said, the pipes to make sure that whatever you kind of capture in that agent experience, you can take that into a different system. And probably in that system, you're going to do the actual sort of agentic operation of evaluating, is this a great lead for my business? Is this the right candidate for my business? We're not building into those things. The one vertical we are building into is training, right? So training is a big use case for us today and essentially learning videos.

8:37And so we are actually building courses, which we're also now from 3.0, which is a way to take, you know, instead of just creating individual videos, you can create kind of a course with many different modules in them. You can add agents. And we're also building a system that actually can check if you've understood the content. So doing comprehension checks automatically. But that's kind of like the only vertical that we're kind of like building deeper in. For all these other use cases, we're essentially going to be this presentation where people sort of put together their experience. But a lot of sort of intelligence and reasoning over the data that you capture inside of the Synthesia experience is going to be taken care of in other systems.

9:12And when you look forward to Synthesia 4.0, 5.0, 6.0, do you see yourself doubling down and tripling down on that video layer, that sort of presentation layer? Or do you see yourself in the future going down deeper into those more verticals as well? I think the other answers will kind of like you know see what happens I think in general it's hard to kind of like predict the future there but I think what we're uniquely good at today is creating very kind of practical utilitarian video right like if you look at most of the content people create with Zodizia people not making advertisements or short films or movies that's not we're not the right platform for that but we're the right platform for it like you want to teach someone something right and that could be when I say training, that's a wide ranging concept.

9:59A lot of people equate that with like compliance videos where you train your sales team, we call it sales development. When you train your customers, we call it customer success or customer support. Like when you train your prospects and why your product is better than the competition, we call that product marketing, right? So it's within all those use cases like communicating information to someone. That's the thing we're uniquely good at. I think that's the only vertical that we'll kind of like continue to build in. And for other use cases, we just like tap into the ecosystem. And one of the stats that you reported recently was that you were trusted by 80 % of the Fortune 100.

10:31Is that primarily in that training use case? It's, yes, but it's a wide ranging training use case, right? So a lot of product marketing is a huge thing. Of course, customer support is huge. Learning and development is also huge, but it is just very much in that kind of like very practical video, right? So another way of thinking about it is like, We don't really help people create video that compete in some kind of a feed. You know, if you're like trying to make social media content or ads, like it's about like grabbing your attention and like getting to emotionally feel something. We're not the right platform for that.

11:05But if you're creating like these like highly practical videos where instead of having a landing page of text explaining how to integrate product X with product Y, that's where you'll see like Synthesia videos. yeah of course but do you see the the way that you're disrupting video right to make video interactive and back and forth do you see that being taken out of the the enterprise world that you're operating in and being applied more broadly to other video formats well so i i do think we're we're going to see that like you we saw open ai release their kind of like um you know tiktok competitor yesterday where they are all that there is like it's trying to innovate on the format where you have this idea of remixing with AI.

11:47I think we'll see lots of that sort of thing where probably your profile can have an AI version that people can interact with and engage with. It's not a world I see us going in. We're very focused on the enterprise. And there's a few reasons for that. I think the first and most obvious one is that in the world of social media, YouTube, TikTok, et cetera, I can't innovate on the format, right? I can help people make MP4 files that can upload to those platforms. but I have zero control over how they innovate on the format. And I'm sure they're going to do something very cool, but it's outside of my control.

12:21And I think it's also going to be different types of use cases you want to enterprise. At enterprise, what we've done over the last five years is building a video platform that spans the lifecycle of the content, right? We help you all the way from generating an idea to actually delivering the video through our video player to the end consumer. And that gives us the kind of license to actually innovate on the video format. the enterprise, they want to host their own videos. They want full control over the videos. They want insights. So I think it's from a practical perspective, right? It's just, it's very difficult to do it in kind of like a down market, a consumer market.

12:55And it's also just about the DNA of the company. You know, I'm a big fan of like focus. And I think other people will be much better than I am. And my team is at figuring out like what the next generation of memes are going to look like. Yeah, yeah. And I think one of the other key stats that I thought was amazing was your NRL. net revenue retention because you know you're very enterprise driven company you've improved your net revenue retention from 119 to 142 how have you done that is that kind of breaking in you know with one product or use case and then upselling across industries and platforms and like how have you managed to improve that to almost like a world-class level yeah i mean it is very world-class and it's even higher when you look at the sort of enterprise true enterprise segments.

13:40I think ultimately, you know, we have this mantra into this called utility over novelty that has guided us, I mean, since we started the company. And it very simply is about like focusing on delivering value for the end consumer, not just creating like novel demos, which sounds very obvious when you build a business. But I think there's a lot of business that falls into the trap of like focusing on what's cool, not necessarily what's useful. And so when I look at a lot of our roadmap, what we've shipped over the last five years, of course, we've shipped like super cool AI things and AI models and the quality gets better, the voice gets better.

14:09But really what we've done more than anything is we've massively increased the speed from someone needing to make a video in the enterprise from an idea to an actual final piece of content. And that really works. And that's not just AI models, right? That's a good editor. It's a good collaboration workflow. It's a content management system. It's all those things. And when you see what's happening with the NRR, it's essentially just that more and more people in the enterprise find more and more use cases is fortunate easier. Five years ago, the quality of the videos and the voices were only good enough for internal low stakes content.

14:42So that's where we focused first. Then the quality of the avatars have gotten better and better, which means that we opened up more use cases. Now a lot of people, I think more than half of people of the video is now external facing content. But then you combine that with the fact that we've also taken the effort into actually making a video that looks good down by maybe a factor of 10X. And all of a sudden now you have not like hundreds of people in the enterprise that can use Synthesia to communicate and create video. Now you have thousands, right? And as soon as you get up into that scale, what you're seeing is just these expansions within the enterprise that goes, you know, 10, 20, 30x of what the initial contract size is.

15:17And that's a pattern we're seeing across the user base. It really is just mostly about like more creators being onboarded and they get onboarded because of better quality of video and faster speed to execution essentially. And we've seen that reflected in your revenue. You hit$100 million of revenue. I think it was April this year, 100 million ARR. And then that was doubling. That was a doubling year over year. Have you seen that level of growth maintained since April? Has it got faster? Has it got slower? What's happened since then? Yeah, I mean, the business is still ripping. And I think with all the stuff for shipping, it's such an early industry still, right?

15:51Like there's a lot of progress been made the last five years, but either of you isolated or something as simple as like looking at the quality of the AI videos, they're getting really good, but they're still not. I still think we have what I call the kind of chat GPT moment of video in front of us. I think maybe with Zara, we're kind of slowly getting there at least for like consumer grade. But once this stuff is truly just indistinguishable from the real thing and as importantly, as controllable as using Synthesia Apertise model today, you're just going to see that. I think what we're going to see is that a very, very large chunk of all the communication we consume on a daily basis is going to move from being text to being video.

16:29So how big do you think this opportunity is for Synthesia? I think we can build a$100 company if we execute correctly in the coming years. And that's taking into account both the kind of core video product that we serve today, which really, the way I think of this is like PowerPoint 2.0. It's actually much less about video because most of what we're doing today is not taking videos that people otherwise would have shot with a camera and replacing the camera with AI. But we're actually doing much more. is taking someone's slide deck or someone's like blog article and turning that into a video. But I think that core product, you know, can get very big.

17:05And then we will layer in this kind of agents on top of it, where actually we can, as a company, go in and create a media format that we own end to end. It's a very different scale of business than what I look at most AI video companies, which is essentially clip generators, right? Like you go in, you type something in a text box, you get a clip out. You can definitely make a big business of that. But I don't think you make a generational business of just like being essentially a clip generator. Yeah, you're transforming communication within all our companies, which is an absolutely massive opportunity.

17:37I want to talk a bit more about the market and AI more broadly. You mentioned that you've said that most AI today is smoke and mirrors and that 5 % truly delivers and transforms industries. is there a way that you are able to see which companies are in that five percent or so is it industries is a application layer model what do you think what makes the five percent the five percent ultimately i think is the nr of retention right i think we're we're in this sort of strange chapter of the market where there's so much fear all up and down the chain of command right the boardroom is where that a new cool startup is going to come in and just wipe the company off its feet because they have some new cool AI thingy.

18:21The rank and file employees are worried that AI is going to take their job, right? They should work in support. The managers of those people are afraid that they're going to be seen as not being forward-leaning enough into AI. So you have all this fear happening in these companies, right? And so what's happening is that everybody is buying tons of AI things. Everybody wants to be an expert. Everybody is very focused on bringing AI into their companies. And I think fundamentally, that's much better than the opposite of being, you know, sort of like very conservative. But what it means is that people are very quick to buy things.

18:53Companies are also very quick to sell things. So there's lots of like iterations happening. There's lots of things being thrown against the wall. And so it's less impressive today, actually, I think to like close a deal than it is to retain a deal 12 months later, because that's where most companies are struggling, right? Like you can close a lot of customers on the promise that your thing will do X, but a lot of these things then ends up not actually delivering. Again, I think it's, I actually think companies are doing the right thing by just trying out lots of different things to figure out what works because it's really like one of those technologies where you need to get like very tactile with it and just try it and do it and give it to your employees and figure out what works and what doesn't.

19:32But I think this is definitely, this is definitely a thing, right? And I think when you just look at a lot of the tools that are out there, I mean, you have the obvious ones, of course, a lot of the big foundation model companies, right? they're absolutely ripping clearly like chat gpd those kind of tools are huge for like people's productivity um and then you have a whole slew of tools where i think uh i i think kind of are overselling what they're doing like if i had to pick on one thing i would say the ai sdr is like a classic thing where i don't know how many like thousands of companies are trying to build like an ai sdr and promising that you can uh you can just switch to this solution for for all your kind of sales teams right and none of them really work i'm sure at one point they are going to work and they're going to be a lot better than they are today but the amount of like chat gpt slop i receive in my email today is insane right it doesn't really work because it's it's just very obvious that it's chat gpt slop personalization is not just like writing an email line it goes way further than that and so i think a lot of these technology companies they end up being very technology centric and not understanding what what problem they're actually trying to solve and so the product ends up like not really working that well.

20:39And so I think as I said, 80 % of them don't work. 15 % kind of stumble to get there. And I think with that, if you can improve fast enough, maybe you can save it. Then 5 % of those companies do phenomenally well, right? Yeah, yeah, no, it's interesting. And yeah, we're seeing more and more companies be very public about revenue milestones. And I think we're starting to see people transition to being more retention focused. And now that it's become the hot metric and that everybody wants to be able to say that they have great retention. We're seeing more and more companies do it. But I think you're right.

21:09NRI is such an amazing indicator of the stickiness of your business. I want to ask two more questions. Number one is just about the year that St. Deja has had in general. You were valued at$2.1 billion earlier this year. You hit$100 million. You've said that you're still ripping. At the same time, the market has just almost lost its head. Things are crazy hot. How is that changing the way that you're thinking about fundraising, the business itself are you thinking okay now's a an amazing opportunity to think about fundraising selling secondaries are you thinking uh we you raised a big round early this year you need to like rest and grow and focus on the cash that you've got how are you thinking about the future of synthasia given the market that is this right now i think obviously we've been doing this for like almost nine years you go up and down in cycles and there's bad times and good times and And right now is definitely like a very fraught of time.

22:02We don't need capital, right? We are very well capitalized. I think basically all the last we've done, we've done kind of optimistically, which is a fantastic place to be in. So our focus is mainly just on maintaining, like delivering value to our customers less than it's about fundraising per se. I do think you're right that like people are talking more about revenue externally because the war for talent is immense. And that is one of the indicators like how well a company is doing. I think things like doing employee secondaries is also like, is an amazing benefit. And I really love to see that happening also in Europe.

22:3711 Labs did a round recently. And of course, part of being public about that is because it's a great place to work, right? It's not just like paper shares you're never going to be able to get rid of. In Suddisi, we've had liquidity events the last two years. I'm pretty sure the same thing is in 11 Labs. And it's essentially a way to attract and retain talent, right? but that said you know there's no no plans to like fundraise I think you know people kind of take things as they come but uh but I think it's uh it's definitely like a frothy market right now for sure yeah definitely um final question you've spoken a bit about UK public markets I think you said some things uh uh what is it um an event over in summer and you know I think in the last 10 days we've had the most positive news that we've had it feels to me that we've we've gone so far down that surely the only way is up Revolut's talking about doing dual listing firmly the big company did a dual listing last week we're seeing two planned ipos of non-tech companies but still you know companies ipoing ipo stamp duty holiday potentially on the cards do you think that's enough do you do you see a change is there momentum that could kind of lift the uk capital markets up or do you think we've still got a long long way to go i i think those are great news i think in particular i've been very vocal in the past about the the sort of uh the stamp duty which I think is just absolutely ridiculous, right?

23:54Like that is so uncompetitive with the US markets. So I think it's great. I think it's still not as attractive as listing in the US, but it definitely feels like we're taking steps in the right direction. The big question is still going to be, I mean, we still haven't really seen any success stories, right? And I hope we'll see some dual listings and I hope we'll see like that model work for Revolut and others. but I think before we can say definitively that it's as attractive to list in Europe as it is to list in the US or at least competitive, I think we need to see the success stories first, right?

24:31So I think a lot of the groundwork is great. Ultimately what we need to see is like, what is the liquidity like? What's the investor interest like? The stamp duty is helpful but if you look at where the pension funds are allocating their money, they still are not big fans of what is perceived as being risky tech stocks. But I'm super happy to see it improving and going in the right direction. And are you thinking about your own timeline to IPO at any point, or do you still think that's too far away? I mean, it's not within the next couple of years, right? We're getting ready for it, and we're thinking about it.

25:04And as you know, it's a process that takes a really long time, but the ambition is definitely to IPO. Love it. Well, Victor, thank you so much for joining me. It's been great to chat. I'm a huge fan of what you're building. I loved your Synthesia 3.0 launch. But yeah, thank you so much for joining me and let's stay in touch. Thanks, man. Likewise. Great chatting. Thanks. Bye-bye.