🔔 Subscribe for more startup strategy and founder stories: https://www.youtube.com/@TrustVanta

👉 Follow us on all social media platforms: LinkedIn, X, Instagram, and TikTok

🎧 Follow Frameworks for Growth on Spotify and Apple Podcasts

‍

In this episode of Frameworks for Growth, Vanta CEO Christina Cacioppo sits down with Steffen Tjerrild, Co-founder and COO/CFO of Synthesia, to talk about what it takes to scale one of the UK’s fastest-growing AI companies.

‍

They explore the future of AI-generated video, how Synthesia built category-defining technology, and why European values may shape the next chapter of AI development.

‍

🔑 Topics covered:

‍

00:30 - What Synthesia does and why enterprises use it  

01:35 - AI video, LLMs, and what's next for Synthesia  

02:57 - The original vision: 2016 research roots  

04:12 - What others missed about AI video  

05:12 - Scaling content creation beyond Hollywood  

06:22 - When the market caught on  

07:36 - Defining the COO role at Synthesia  

08:47 - Building a services team to accelerate adoption  

10:56 - How customer insights shape product  

12:39 - Synthesia’s customer-first evolution  

13:43 - The highs and lows of creating a new category  

15:31 - Building a mission-driven company  

16:09 - How Synthesia’s mission evolved  

17:33 - AI ethics, content control, and platform trust  

18:31 - Synthesia’s AI safety framework: consent, control, collaboration  

21:42 - Steffen’s regulatory wishlist  

22:58 - Enabling brand-safe, aligned video at scale  

24:00 - What we’re not talking enough about in AI  

25:44 - Building in Europe and retaining top AI talent  

27:49 - Pitching Americans on working in Europe  

28:27 - Why Synthesia builds its own models  

30:10 - How custom datasets create differentiation  

31:48 - Avoiding novelty and prioritizing utility  

33:11 - Finding product-market fit  

34:05 - Breaking old paradigms of video production  

35:33 - Lessons from startups in Zambia  

37:10 - Using Synthesia in underserved regions  

37:58 - Advice for founders in the "idea desert"  

38:46 - Why the company didn’t die  

39:17 - Founders should invest their own money first  

41:03 - What Steffen looks for when angel investing  

42:10 - Manual habits that keep Steffen grounded  

42:47 - What he’s reading now and lessons from it  

43:40 - Operator role models: Stripe’s founders  

44:17 - Why 100% retention isn't always the goal  

45:57 - Key operating metrics and evolving go-to-market  

47:27 - What felt huge but wasn’t  

48:38 - What seemed small but mattered  

‍

Read the full transcript 

Christina Cacioppo:

Welcome to Frameworks for Growth, where we explore the systems, frameworks and tools that the world's leading founders, operators, and investors use to build and scale. I'm Christina Cacioppo, co-founder and CEO of Vanta, and I'm here today with Synthesias COO and co-founder Steffen Tjerrild, thanks for being here today.

‍

Steffen Tjerrild:

Thank you very much for having me.

‍

Christina Cacioppo:

So for those unfamiliar, could you tell us a little bit about Synthesia?

‍

Steffen Tjerrild:

Sure. Yeah. So Synthesia is now the world's largest AI video platform for enterprises. We allow customers to kind produce videos with computers and not with cameras and microphones and actors, but actually taking video creation into the 21st century by allowing companies to produce content behind their desk. And that is just a very incredible way for companies to scale their video production. I think we all live in this kind of video first world. Everything we do in our private lives is very media and very rich media. But when we come into work, especially in the biggest enterprise in the world today, right, it's very much PDFs, PowerPoints, emails, word documents, and there's just a starer and starer difference between how much content they want to produce and how much they actually produce. And also as the native generation coming into the workforce that is demanding more interactive and more engaging forms of communication that is just becoming the norm and companies are craving new ways to keep up with that kind of content needs.

‍

Christina Cacioppo:

Very neat. And you're at the forefront of AI video production. What are some things you're excited about in tech today?

‍

Steffen Tjerrild:

I think there's so many very exciting things, and especially with LLMs are just incredibly magical technology. And I think there's going to be a forest of incredible new small companies and companies being produced out of this. And I think we are really excited to just take this new level of reasoning that computers are starting to grasp and really crafting that into new experiences with Synthesia. So it's just an incredible kind of time to be alive. And I think also with the scale and resources that we have as a company, I think we are in full position to leverage a lot of these kind of breakthroughs. And I think for us, we've been going at the company for many years now, but we still feel we are 10 or 15% into the roadmap of creating Hollywood style kind of videos on our laptop, which was our initial tagline many, many years ago. And we used to say that I was 10 years away. I think we're actually going to be more close to that in the coming years. And so we were pretty right on that prediction. But yeah, I think there's just incredible opportunity ahead, especially as LLMs become incredibly magical to allow new techniques and workflows and experiences to come to life.

‍

Christina Cacioppo:

The idea of text to video sort of wasn't a thing before you all, but you all believed in it. What did you see?

‍

Steffen Tjerrild:

I think we probably have raised a lot more money early on in our journey if we had our crafted vision for where this category was morphing into. But I think we had at least the crazy kind of moment when we saw this technologies for the first time back in the fall of 2016, where the professor, one of the co-founders in Synthesia called Professor Mattias Neener for the first time showcased a paper called Face-to-Face, which was the precursor of these technologies that was coming on the back of the VR hype cycle back in 2016 where a lot of deep learning computer graphics, computer vision really coming together to enable some of these things. And when Victor and myself saw this, it was an incredible kind of glimpse into the future of content creation. I don't think we had the foresight to see the product expression that it has today, but it was at least crazy enough for us to quit our jobs and pursue this idea.

‍

Christina Cacioppo:

And so you all saw this but kind of everyone else missed it. I'm sort of what we're saying, but I'm sort of not. What did everybody else miss that you all understood?

‍

Steffen Tjerrild:

I'm not sure we kind of understood everything. I think it's easy to kind of draw all the lines and the dots in the back rear mirror, but I think what we saw and what we had conviction about and what people had misunderstood is that there is just a long tail of people that want to create video but can't do it today. I think the biggest invention in video production has been our cameras that we have on our phone, but it's still inherently physical. You have to put that up, you still have actors and studios and these type of things and actually digitizing that motion a bit similar to how other creative disciplines has kind of been disrupted by technology. If you go back 20 years, you need to record everything. And for music production that used to be in recording in a $10 million LA recording studio today, you can synthesize any tune the universe on your laptop and every music and laptop chatting hit today is synthesized with the computers and not with physical instruments.

‍

And we just saw that of course, video is going to be on the same path and video is going to allow us to produce content behind our laptops and less with a smarter or faster kind of physical production process. And I think this starting point with avatars, that felt pretty odd because if you're speaking to investors and consumers and customers, they're thinking, oh, did look janky or robotic and these type of things. But if you compare it to text, that is where you have to compare it. You don't have to compare it with hundreds of millions of dollars in visual effects budgets from Hollywood. And that is the market net new kind of video creators versus existing creators. And that's also a little bit of the attraction we have in the companies today, right? It's less about making the video department 10 times more effective. It is to make all the consultants and your consultancy video producers, maybe not for every PowerPoint but for every 10th PowerPoint, maybe make a video because it's more engaging, it's more intuitive to disseminate information through video.

‍

Christina Cacioppo:

And you mentioned clouds over the category. When did it feel like those lifted a bit or when did other people start to get it more?

‍

Steffen Tjerrild:

I think we were actually growing really fast even before chat. So we launched the first version of the product back in 2020 in summer of 2020. And we had a lot of acceleration there. People couldn't hire crews, people couldn't actually do video production, but we are actually growing really fast. And I think after chat GBT landed in early 23, we just saw a huge acceleration. AI became on everybody's lips, everybody had to do things with ai and Synthesia was just in a very great position that here's five things beyond ChatGPT you can do, and Syntheses was one of them. So we had just had this crazy kind of lift off in terms of the pull in the market and it also becomes much more accepted that, oh, using avatars, using videos in your customer communication, in your internal training so forth was just becoming much more of a standard back then. Yeah.

‍

Christina Cacioppo:

And at Synthesia your COO to the role, I think probably it can be so many things or it tend to be a little misunderstood, I think. Is there a decision you made recently that was really impactful that people outside the company wouldn't know about and not going to use this to try to understand what does COO mean?

‍

Steffen Tjerrild:

Yeah, I think CEO in Synthesia context is probably all the rest. It's encapsulating a lot of the things delineate between Victor, my other co-founder and myself is that he's more product r and d typically painting the story with the outside world where I'm typically behind the camera and sitting with operations, sales and finance has kind of been typically running into me where Victor thinks about the future. I'm thinking about the present and reporting a bit on the past, a recent example, I don't know if it's a good one, but a thing we had this idea, okay, Synthesia is super simple tool, it's very inviting, it's designed for people who has never created videos before. So it's kind of by purpose, pretty easy and inviting tool. But we realized that even though that it might be easy for tech natives, when we are selling into the biggest companies in the world, that's typically relying on third parties, consultants, agencies to kind of do a lot of the execution and process orientation in the big companies, we kind launched and implemented a lot of services even though the tool is relatively simple, we saw that actually going in and helping our customers crafting the first video library for some of those use cases.

‍

So implementing our services team was a decision we made that I think is really starting to pay off where we have template designers going in and taking a customer's brand guidelines and creating video assets that allows them to just find value quicker. Where I think we tech nerds are just very much like, oh, it's just go in, it's super easy, you'll figure it out yourself. But actually having the appreciation for primarily a senior work professional in the enterprise doesn't have the same attitude to just jump in and pick new technologies and new techniques up.

‍

Christina Cacioppo:

How did you build out that team? What sort of folks did you hire? What were you looking for?

‍

Steffen Tjerrild:

Yeah, so we are looking at other technology companies that has been, I know more technical than ours and I know how do you actually craft that? Can I know services arm that really allows the company? And here it's less about building a profit center, it's not about that, oh, we're going to sell services and going to be I a strong kind of p and l line item. For us it's more about how can we set this account up for long-term success and less about like, okay, let's maximize some shortterm willingness to pay for implementation services, but how can we actually set this up account up for not even just in year one, but in year three and five, if we can accelerate those learnings over six to nine months instead of the customer, we'll eventually find those over 18 or 24. I think that was a little bit okay, this kind of clicks doing things that doesn't scale.

‍

And also for us in such a new category with customers that want to pull the product in many different directions, actually allowing them to, oh, you'll figure it out with our API actually going a bit deeper with our customers and trying to understand, okay, so what is the integrations you want to make? What is the last mile distribution we want of these videos? And actually going a little bit deeper with our customers and building with them to enable these kind I know net new use cases and net new experiences that they have not been doing before.

‍

Christina Cacioppo:

And did that these kind of net new experiences your customers build, how do you get that information back to your product teams?

‍

Steffen Tjerrild:

Yeah, so we are building, we working, we are lucky to having a lot of big customers and we work with 70% of fortune hundred now, which is an amazing kind of position to be in. And I think just to kind of back up before I'm answering that question, we've kind of been through different phases as a company where I think the first years of the company where our research first company can we actually build a technology where we can type in text and we'll get a person to say that that was the first milestone and can we actually make a scalable offering around that? And once we broke through and actually had the technical breakthrough to allow this, we had to build a product around it. Our next realization was, okay, there's more to video than just somebody talking to you. There is animations, transitions, music, all these things that actually makes a video engaging.

‍

So then we build a full video editor around the technology that allow customers to build end-to-end videos inside Synthesia instead of exporting to other kind of video apps. And that allowed us to build a really strong application that people like love today. And that was just an incredible long roadmap to get to feature parody with a common kind of video editor in the cloud. You can imagine the long list of features with drop shadows and animations and custom font and all these things. And then the last chapter that we're entering now is being much more customer first sounds odd as a company being customer first after seven years. But I think now we actually, the biggest mode and the biggest advantage we have is actually the distribution, the trust that we've built in the enterprise and actually allowing us to sit at the table and allow us customers.

‍

They are asking us how can we solve this problem with AI video? How can we solve that problem with AI video? And I think it's just giving us unique position to allow us to build with these customers and leveraging that with our services team to go that extra mile, understanding the nuances of that use case and that kind of application. And the way we feed it back is on an aggregate level, it's a really big initiative for the company. So actually also sitting with the entire C-suite from product to our CRO to CEO and reviewing, oh, this company want to do this, they're also trying to find, there's a lot of nuances that customers want to have a very specific things and trying to kind of dissect is this a customer a kind of situation or is this actually something that's more broadly? And obviously that's also the hard questions.

‍

Christina Cacioppo:

And as you're doing this right and you're talking through it, you're defining a new category and that sounds so cool, especially on the outside the stuff of founder dreams, but you're actually doing it. How cool is it actually? Like pros, cons.

‍

Steffen Tjerrild:

It is a blessing and a curse. I think when it works, it's obviously amazing and I think that's also why all the founders of here and a lot of the early employees are still here is that you're kind of riding the wave of this kind of white space. You're not like rip and replace existing vendors, you're creating a knit new market, which is obviously amazing, but it's also a challenge because there's not currently a line item in most companies as AI video. So there is obviously a lot of trade-offs here, but I think it's exciting because it's also an exciting technology. I think even though that the technicalities and the deep AI that we build behind it to enable these things is very complex. The actual output is very simple. David Beckham can't speak Japanese and allowing him to do that, even my mom understands that.

‍

And I think that's a very cool part of the technology that we can express something very simple and so intuitive that most people understand or can type in text and get this person to say that and create a video around it. That's crazy. And that's the beautiful thing. And I think it's also coming a little bit back to if I had to start a company again, I think I love this saying that it's easy to build a hard company and it's hard to build an easy company in the sense that when you're actually setting out to build the Hollywood style movies on your laptop, then you also get a very different kind of talent pool that want to be a part of that mission, want to be a part of that journey. And basically as two younger business co-founders at the time with no credentials in AI at all, I think it helped us to actually have some sort of an offering to these amazing people that could go and work at Fang, right and get triple the compensation that we could offer them at the time.

‍

But actually having a pretty crazy idea of guys we're going to try to make Hollywood style movies on our laptop that allowed us to tap into a different talent pool and get more missionaries along that actually was here for the right reasons and want to know was inspired by that. Versus if you're building accounting software number 500 incrementally with maybe some ai, you'll probably get more of a transactional employer employee relationship and people just coming for this paycheck. It's not like that there's a crazy mission or objectives that you're trying to set out to do.

‍

Christina Cacioppo:

And when you talk about the crazy mission of an objective of Synthesia, how do you talk about it?

‍

Steffen Tjerrild:

I think it's evolving a lot over the years and I think there might be some companies where the mission is just very enduring through the entire, but I think also as AI just enabled more and more things that you couldn't even think about a few years ago, I think we are kind of morphing more into, we started with this kind of Hollywood style innovation back when we started the company. I think we morphed in much more business knowledge and actually sitting and helping customers being much more effective in their communication. And then knowledge sharing if that is internal or it's external with the customers, right? Information is just getting harder and harder to disseminate, especially when the tension spans is going like this. Having the ability to allow our customers to drive higher and more effective communication in all that channels is kind of much more the kind of direction of travel that we're going in.

‍

Christina Cacioppo:

When people talk about AI, there's obviously a lot of excitement and then they're going to challenges that are more specific to ai. So things I've heard are worrying about DeepFakes or accuracy of the models like David Beckham speaking Japanese, great David Beckham maybe, I don't know, yelling at your grandmother with curse words less great. No one wants that. How do you prevent that? And then people also talk about just the cost and the cost of goods and the financing required. What's another challenge of building an AI that maybe we talk about a little bit less?

‍

Steffen Tjerrild:

I think definitely the ethics side of things. I think AI is a different type of vendor than a traditional kind of software vendor from historic process, both from security, the ethics, the kind of principles that they're training, the models and so forth. So this is something Synthesia has been thinking a lot about and incorporating into the company since day one. So that is something that we think all the customers and companies will think even more about in the future as it becomes more prevalent and people will care about who they partner with, especially in this era of I know where data and the gold that companies and the IP that they hold is being more challenged by a lot of the kind of internship that we see.

‍

Christina Cacioppo:

And actually to that, you all have an AI framework, if I'm correct, with consent, control, collaboration. How does that manifest? What's a specific decision you made with that framework?

‍

Steffen Tjerrild:

We had the framework in place from day one of the company and it was very important for us to set the standard of this. We knew these technologies was incredible controversial when we started out. So we wanted to build a framework that allowed our customers, our partners to feel that we are in good hands and we take this very seriously American just to go through the different Cs. I think consent, there was no regulation at the time, there was no laws that says we couldn't do anything, but we were just like, okay, it probably makes sense that you can't just take somebody's face and make them say something they haven't said before. So we made this framework explicit consent. So all actors, all people on the platform, all personal avatars have to go through and give explicit consent to the fact that they'll be part of the strategic platform.

‍

And here a lot of our competitors made maybe fee cheap celebrities, not necessarily maliciously, but for fun or for awareness or for PR and so forth. Actually just taking a politician or a celebrity and make them say something funny. And we've just never done that because we had a pretty hard line around, it has to give explicit consent and the amount of times over the years where it's very benign, but people want to do a video with a deceased actor or something even for a good cause. We just always just said no. Even though that it might have made some short-term awareness and pr, but we've just always said no right to that. And I think the next one around control. We are also pretty harsh in content moderation, which is making sure our platform is not being used for bad. And I think here it's actually kind of the red content is pretty easy to filter out profanity, hate speech, these type of things are actually pretty easy.

‍

There's more like the great content, which is the hard part. So we actually employ, a big part of the company is human moderators that is actually moderating the content. And here I'm glad we are not like a free speech platform or something where we have to have very liberal guidelines around this. We can be very strict. We work with enterprises and our enterprise customers also want to make sure that their employees is not creating content that is not on their brand guidelines. And so far, so this is actually turning into competitive advantage as well that companies are feeling more safe, that they're in safer hands, that we are protecting their brand when they're using Synthesia to create video and collaboration. I think one thing is Synthesia being not misused, another thing is internet scale. How is these technologies being leveraged for good and how's the harm minimized? So also been doing a lot of work with governments and media organizations on awareness, but also the big tech giants where how can we actually get to a media reality in a few years where we can kind of trust the content.

‍

Christina Cacioppo:

And to that there was nothing, no regulation, nothing when you started, if you were master of the universe and had a magic wand and there was one guideline even you could wave your wand and make happen.

‍

Steffen Tjerrild:

I think our kind of framework is kind of pretty simple. It's trying to use common sense around these technologies and that's not to be too political, but I think that's how most regulation should be. That's just kind of coming down to sensible common sense. Of course you should not be able to take somebody's face without their consent and make them say things they shouldn't say. And I think some sort of standard upholding of content moderation around the standard free speech principles that's also governing. You can't just say everything just because there's kind of free speech, just principles around that is also interesting. And I think the way we are taking control to the next level, you can see Synthesia as a canvas and you can write pretty bad things inside Word or even in PowerPoint. We are actually moderating it at the creation level where all moderation today happens at the distribution layer, which is on the distribution platforms, but we actually doing it before and it gets any distribution at all. And that's obviously a next even further step to make sure the platform is extra secure.

‍

Christina Cacioppo:

And do you do that in partnership with your customers? Is there a level you do outside of even what a customer might have explicitly told you? How do you think about that?

‍

Steffen Tjerrild:

Yeah, it is also in the LLM is changing this a lot, but we are pretty excited about also allowing our customers to control this a bit more on their side as their LLMs is becoming more sophisticated. And one thing is on the content moderation per se, but if you're a big company you want to make sure your employees is talking the same way about the product. That could also be governed on the moderation layer. So when you're trying to talk about this product, hey, this is actually not how we talk about it, this is the standard phrasing or the standard acronym for this business unit or whatnot, allowing the companies to unify the communication and share the same language, which just becomes extremely hard at scale. And that is a lot of the kind of customer feedback we see today, right? A customer's like, yeah, it's great, but how can we make sure that we speak the same language around the world? How do we have the best practices around the world in a more kind of uniform manner?

‍

Christina Cacioppo:

And you're sitting at the intersection of content, AI, identity, seeing a lot. What's something the rest of us aren't talking about as much as we should be?

‍

Steffen Tjerrild:

That's a good question. I think there's been happening a lot on security and that kind of path I think is definitely evolving and people are taking more that serious of who do you share your data with? One thing is I know what is their security and compliance controls, but who is the owners of the company? Where do they come from? All these things is a next level deeper of how you're making sure that you're in bed with partners that you want to build a relationship with. So I think that's increasing. I think just generally AI is obviously on everybody's lips, but I think people also sometimes inflating or jumping to the end state. I think most people can imagine the perfect chatbot that just answers your questions exactly how you want it or retrieves the exact information that you're asking for. But I think reality is just an humans and companies and businesses are just still very messy. Data is just very siloed, it's very disorganized still. And I still think that we need to sequence our ways to a lot of these magical outcomes that everybody can imagine, but actually realizing there is a lot of our why today, but how can we actually build a sequence to get there? And it's hard to sometimes just jump the fence to the end state. You have to actually build a lot of things to be able to get there eventually.

‍

Christina Cacioppo:

Synthesia is built in Europe and as we're talking in 2025 seems to be more likely to be a forefront for AI regulation in the US. Building in Europe is a helpful constraint. Something else. How do you think about it?

‍

Steffen Tjerrild:

It's not something we are thinking a lot about to be honest. I think obviously we want to be in an environment where it's encouraged to know build big AI companies and that kind of coming with some sort of frameworks that we need. But I think the UK is out of the EU now and it's kind of taking a slight different stance to the rest of Europe. But I think our kind of view of the world is that there's amazing people in Europe and very ambitious people as well. I think there's less that have seen the last mile on that really big scale that a lot of the west coast companies have, but I think the ambition level and the talent density and I think also the talent loyalty in Europe is relatively higher, all things equal. So I think we have been generally happy and I think especially the first few years of the company, that was just a really, really rough patch. If we were a standard company in the valley and there was five other cool things around the corner, who knows, right? I think we'll probably have a hard time retaining those great AI people back then that allowed us to actually make those breakthroughs in the beginning. So I think we're kind of generally happy. Of course we want a environment that encourage us to build and stay here and I think we are kind of well supported so far.

‍

Christina Cacioppo:

Last night actually even at a company event, we were talking with Europeans who moved to the US for a few years, but they really want to live in Europe. And so having a set of companies they can join and feel they're both pushing their career forward and living where they want to does feel like a neat talent advantage.

‍

Steffen Tjerrild:

Yeah, totally. And we've also been lucky to attract some great operators from America that's migrating over here the other way. So that's also been helpful to allow us to have people and executives that are seen that really, really big scale, which is exciting.

‍

Christina Cacioppo:

And what's your pitch to Americans to come over here?

‍

Steffen Tjerrild:

I think Europe has a lot to offer and I think it's a cool, cool experience and I think I know the cities are amazing in Europe. I think that's generally something that attracts a lot of American. I think London and Paris is also very idolized, so my wife is American so I've been able to convince her to come over here as well and she's never going to come back, but yeah.

‍

Christina Cacioppo:

Well you all build your own models, correct with that. Where do you think about differentiation?

‍

Steffen Tjerrild:

I think it is a good question and something we asking ourselves on a constant level, but I think we see it ultimately as a kind go-to market advantage in terms of time and scale where we also have a very specific kind of customer and outcome in mind and it's less just a foundational model that has to cater for everything. We have a very specific user and a very specific use case in mind. So we also very much building for that kind of particular workflow and those constraints that those kind of use cases have. And I think we're actually pretty pragmatic about it. I think a few years ago there was nobody else that actually had avatars, we invented that category. So we're actually enabling technology for us to actually have a claim to fame and building the company around that advantage. I think relatively that advantage is going down as can I know all kind of technologies getting more and more commodified over time.

‍

So we see ourselves as less of a foundational company, more like an applied company. And I think eventually when big companies or open source comes out with crazy kind of new things, we will get it into a spot eventually where that will just be something we integrate into our user experience. Our customers is not buying an AI model from us, they're buying a workflow, they're building an experience, they're building a workflow tool that works for them to address a particular use case and Avatar is just an ingredient of that, but actually the user experience, the integrations, the lab mile support and distribution of the video and enabling net near integrations and experiences with video is kind of where the value is and what the customer is buying from us.

‍

Christina Cacioppo:

A lot of folks now talk about data as a competitive advantage in ai. What's your take on that?

‍

Steffen Tjerrild:

Yeah, I think so. I think especially I think for AI and for language models or voice models and stuff like that, I think the data is definitely expressing the model capabilities to an extent. And I think that is definitely going to be where incumbents is going to have a big advantage over newcomers because they sit on a lot of proprietary data that is not just online scrapable that a lot of these LLM companies has been doing. So I think that is definitely something we see and that's also something we are doing a lot about where to my point earlier, we have very specific use cases. We care about human presenters in video, so that is a very narrow, at least the confined domain of content that we need to specialize and be the best at in the world. So curating buying, we just did a deal with Shutterstock where we're buying a lot of their content, allowing us to train models on this. We have been hiring Broadway actors to know what does a great performance look like in video. We had a $5 million, I know volumetric capture studio here in London creating the highest visibility data set for how does humans perform in video. So we've kind of been doing a lot to kind of curate and build the best dataset for this in the world. And I think that will be a differentiators piece in the short to the medium time.

‍

Christina Cacioppo:

Brilliant and a lot of AI and especially when you think about characters that could lean toward novelty or could feel really in a novelty camp, how much of that do you want to avoid? How do you avoid it if you do?

‍

Steffen Tjerrild:

I think the first few years of the company, which was very much going through the product market fit desert, really allowed us to have the conviction when we actually found something that worked to not get carried away with these kind of innovation potholes or getting too carried away with cool and novelty ideas. And I think we've had this saying as a company that utility of a novelty and just been extremely focused on the utility that we as a company are providing today, not tomorrow, not in six months, not in two years, what is the absolute highest value of the product today? And that has kind of been guiding our GoTo market efforts and our focus where I think it's very easy to get carried away, especially if you're a young founder and just trying to get some traction or find some customer that want to buy from you. And I think those first three years of going in through the idea maze and really trying many different different things and failing a lot of different things just gave us that absolute conviction. Okay, let's focus on this. Let's double down on this. We have a customer base that loves us and not get carried too much away. And I think that's all about sequencing this and also when is the technology next gate? What is the fringes of the technology that can be the next TAM or expansion or the next use cases for these technologies.

‍

Christina Cacioppo:

When you were wandering through the desert and the idea maze and when did you know you had something?

‍

Steffen Tjerrild:

It was pretty instant. I think once we actually just launched the self service offering around there was just instant product market fit and I think we obviously had no idea who's going to use this and who's going to be things. We obviously had some hypothesis, but it was just very clear that there was this long tail of people that has never had the means or the ability to create content before and that was just incredible to find that and double down on that and building for that, listening to them to refine the offering. And I think the first thing when we launched the product, it was a avatar clip generator. You can kind of generate a clip with an avatar, but then the next was, okay, let's build the editor around it that sits around that. And I think one of the other interesting things in product decisions we have made or many product decisions is around the paradigm for generated media is very different than camera recorded content.

‍

So we had just been building and making a lot of product decision, what does it actually mean when there's not a physical timeline, when the content is not recorded from A to B when it's not physical kind of content that is with physical sensors and we already have all the amazing infrastructure to work with camera recorded content and also the amazing distribution online and hosting capabilities and so forth. But what does it actually mean when every pixel in the video software rendered as opposed to sensor captured? And that has allowed us to build a lot of unique innovations around how do we work with this when you're not confined by the old paradigm? And that's also why we had the conviction that this could actually get pretty big because a lot of the incumbents of camera recorded content are still working in that paradigm and all their innovator dilemma around that. And I think that's kind where we are so bullish that we actually have an air gap among the big companies to kind of build a really big category around this because we think there's so many new things that can be enabled by these technologies.

‍

Christina Cacioppo:

So you studied business in Denmark, moved to Zambia, started some businesses in mining and logistics and chickens. I have it right. We've also spent some time in the west coast of the US but what'd you learn about entrepreneurship with Zambian chickens at the west coast? Kind of misses.

‍

Steffen Tjerrild:

I think. Yeah, maybe just back up a little bit.

‍

You want to go straight to the Zambian chickens?

‍

Yeah, so coming from Denmark, used to school there, study in the US as well, but having also a pretty big foothold and know the Danish startup ecosystem also when no Victor, my co-founder, we started in this incubator that starts companies. We learned a lot what not to do when you're building companies. So there was some great learnings there, but we saw that we were complimenting each other really well. Our Venn diagram was pretty low in the overlap. We had very similar grandiose ambition, but we didn't have a lot of the same skillsets. He's more technical and more of an artist and I'm more like the executor and kind of the finance background. But then after I graduated, the most crazy thing I could think of was to go to Africa and I was lucky enough to have an opportunity to go to Zambia work in a regional investment firm down there and I learned a lot about resiliency and the hardship of building companies in very adverse environments. So that was just an incredible experience. It's not too many gener AI companies at the time, but it was definitely incredible experience and two amazing years of my life where I was looking and investing and operating everything to point chicken farms to mining supply chain and microfinance and so forth.

‍

Christina Cacioppo:

And I know this isn't entirely a focus now, but if you were to apply some text to video and AI to your prior life in Zambia, what would you do? What stands out?

‍

Steffen Tjerrild:

I think there's obviously some educational gaps in Africa and I think having video as a medium to bridge some of these literacy rates I think is a great way to potentially leverage the technology to scale up and scale up the general population around different topics and around different principles. That is something we are thinking about and we are working with actually NGOs that is producing content with the platform to some of these kind of regions.

‍

Christina Cacioppo:

And as you were then, so after the chickens came up with the bunch of ideas and we're going to idea maze wandering in the desert, did it have to be that long? Did you have to ship stuff to learn? When you look back on that, what advice would you have for someone in the desert idea desert now?

‍

Steffen Tjerrild:

I think we probably could have done it maybe six to nine months quicker than the three years, but I don't think more than that. It was literally a technical breakthrough that we needed to actually enable the product and we were the only company in the world that could do that at the time. I think if we had that clear focus and that kind of clear end goal initially we could probably have done it faster and work more. I guided towards that outcome. But I think it was a bit just serendipity that we managed to stay alive, which was we were pretty close to death a few times, but I actually managed to stay alive to see the day of life to actually see this kind of product to come to life. Yeah, that was pretty hard.

‍

Christina Cacioppo:

How did you stay alive, why didn't you die?

‍

Steffen Tjerrild:

I think just never give up. I think it was just there too much in line. We were like young people. We were both very young at the time trying to just, we were just living together. So it was a pretty hard time, but I think it was just this fear of letting people down, investors, employees and just trying to go through day by day.

‍

Christina Cacioppo:

Is there, before raising money, is there something you think all founders should do, many founders should do?

‍

Steffen Tjerrild:

I think they should definitely use all their own monies. Yeah, I had the credit on all my credit cards and everything. Until we land the first financing round, I think we see that there's maybe a little bit of a spoiling of the earlier generation that, oh, we can just raise around blah, blah, blah. But actually going through and showing that you are all in, you quit your job, you've spent all your money, you're kind gone to the edges of where you can actually do it. That shows the commitment, the actual knows you are all in and shows that you have the grid, you have actually the hardship that it takes to really do so I think that's a symbol and it's very hard advice to take, but I feel like that probably more should do that.

‍

Christina Cacioppo:

And I both agree to that. But why is that important? Sometimes I have trouble articulating it.

‍

Steffen Tjerrild:

Think it just shows that you are going to do this at all costs because that's what it takes. This is going to be the easiest part to deplete your bank account. The hardest part is to actually get financing, operate a team, building a company that is just extremely hard. And I think if this is already a thing where you don't want to cross the line of what it actually takes, then that's probably, that's an early sign of where are you not willing to compromise your life because it's going to take over everything that you do and it's going to take every waking hour of your life and even your sleeping hours as well. So I think just showing that commitment, understanding that this is what it takes, this is probably a principle, at least a lesson that you have to take on early.

‍

Christina Cacioppo:

Yeah. Is it something you look for when you're investing in Southern Africa consciously or not?

‍

Steffen Tjerrild:

Probably subconsciously. I think it's hard to really test that with those pretty transactional relationships that you have, or at least when it's not people that you truly know. But yeah, I think it is something I probably should think more about when I'm angel investing, at least.

‍

Christina Cacioppo:

And you angel invest now? Anything in particular that you look for?

‍

Steffen Tjerrild:

The crazier the better. I think that's typically where I think incremental type of things could be interesting and can probably create a good business on them, but I think trying to push the humanity and pushing the boundaries of what's possible to the next level I think is what excites me.

‍

Christina Cacioppo:

Yeah, similarly it's like there's a lot of B2B software and it's great. I like B2B software, but I spend a lot of time of my time thinking about it. It's kind of nice to hear about something that's more like asteroid mining on the moon sort of thing. Exactly. Well, you operate on the bleeding edge of ai. What's something in your life you do Totally manually? Smoke signals maybe.

‍

Steffen Tjerrild:

At work or privately?

‍

Christina Cacioppo :

Either way.

‍

Steffen Tjerrild:

I still have my one-on-ones in person, not yet my avatar, my voice agent that is conducting those, but I bike a lot as well. So I think simple things like that is still something that we as humans are getting gratitude of doing stuff, simple things.

‍

Christina Cacioppo:

I know you're a Strava person when you cycle.

‍

Steffen Tjerrild:

Yeah, a little bit, yeah.

‍

Christina Cacioppo:

Favorite segment?

‍

Steffen Tjerrild:

Not yet. I'm not that Strava person.

‍

Christina Cacioppo:‍

What are you reading right now?

‍

Steffen Tjerrild:

Currently I've been reading Bill Walsh Skull Take Care of Itself. The thing, it's kind of coming a lot with my personality of if you just take one step at a time and focusing on the right things, the outcomes and the things will eventually kind of appear, not get too carried away. And I think it's cliche, but I think also the journey is the fun. You have to enjoy the day ins and day outs of doing this and doing it with great people and then eventually if you continue to do the right things and focus and optimizing the right things, the score and whatever KPIs and metrics that you're measuring will eventually go that way as well.

‍

Christina Cacioppo:

Yeah, and when you think of actually that perspective and applied to company building, is there a founder or an operator who you think does that super well?

‍

Steffen Tjerrild:

Big admiration Stripe founders or investors in Synthesia. They've obviously been executing extremely well and doing also with their own kind of principles and they're hard in it as well. And it seems like they've also been obviously incredibly hard, but also kind of been doing it with their own soul in it. I feel like they can be felt through it. That's a good one.

‍

Christina Cacioppo:

Did you have to rewire yourself or folks who work at Synthesia to be like a hundred percent? Retention is not the goal even though we all went through school where a hundred percent was always the goal.

‍

Steffen Tjerrild:

I think it's obviously something you also have with investors and outside people writing, comparing user benchmarks and these type of things, which is obviously helpful and interesting. But I think to my point earlier, if you're actually striving for a hundred percent retention, you're probably not ambition enough and you're not pushing the technology to where the edge is. But I think it's obviously a very critical piece and that is where the true weighing machine of a technology or company is actually is. It's less about the initial buy because there's a lot of loose money, especially in AI today. People are very willing to try stuff. But on your one year renewal, that's typically where the true sentiment, was this just a fun pilot and a test or is this actually meaningfully valuable for me that I want to continue to use and build value on? So I think it is a very important metrics to kind of drive your business around, but I think you can also be too over-focused on it.

‍

And if you are not granular on that churn or retention, I think you can also get too blinded around a headline number that maybe looks lower than standard, but if you granularly look at it, is this like a super core ICP or is this a fringe marketing use case that we've never seen before? We are not going to turn that customer away. We're going to say, Hey, cool, try that out. But we also going to not be as upset or so surprised if that use case maybe not materialized because that's maybe on the fringe of what the tech can do today.

‍

Christina Cacioppo:

And when you think about operating the business, are there a couple of metrics that you pay a lot of attention to?

‍

Steffen Tjerrild:

I think obviously our retention is important, but I think our net retention is extremely important as a business, especially in the biggest companies that we operate with where we are just want to see more and more teams picking it up and want to go more wall to wall with our customers where it is still very much bottoms up technology where we are solving a lot of user pain. We are in the process of kind of maturing from user pain to business pain. Where as a PLG company solving individual user pain, Hey, I used to produce this material with a camera, now we can do it cheaper and faster with AI video, that's great, but attaching that to kind strategic business outcomes that budget holders decision makers are really sitting on, that is kind of journey that we are on and that's some we want to working hard on how to go into more senior people they don't care about if it's video or text or message birds that is delivering the messages, they are caring about their customer support. Tickets are being deflected more because you're now using video versus text. They care about if their product marketing is more effective. So you can better cross-selling the sales enablement team cares about ramp time if you can compress that with video, that is something they care about and so forth.

‍

Christina Cacioppo:

When you think back on the last seven or eight years of synthesia, what's something that felt like it was a really big deal at the time and in retrospect was kind of nothing.

‍

Steffen Tjerrild:

I think funding rounds is often feels like this end all be all moment, but I think it's essentially just a means to know executing on your vision and mission. And I think they're obviously important, but I think they're obviously overstated. I think they're a little bit over hyped in especially for more seniored and seasoned entrepreneurs.

‍

But I think that's probably something that you overstay. They know the importance. Of course you need capital if you're having ambitious goals and need capital to do that. But yeah, I think it's typically overstated, but I think it's also the other way. I think things are rarely as bad as they initially seem and they're rarely as great as they initially seems as well. So I think we as humans are oscillating in these kind of highs and ultra lows all the time, but they actually know is rarely as bad or rarely as great as it initially.

‍

Christina Cacioppo:

Also true then in the moment. Yeah. Is there something though that when it happened again, didn't feel like anything and in retrospect actually an important decision or an outcome or win or loss?

‍

Steffen Tjerrild:

I think when you're losing potentially big customers, that obviously happens when you're forming a new category and these type of things that can feel like a huge blow and the company is over and we're never going to win that logo back or whatever. I think in the grand scheme of things, this is also just the nature of building a category and trying to take the technology to the edge of what's possible. I think if you have a hundred percent retention in a category, you're probably not ambitious enough, especially when you're kind of forming a new thing. But in the grand scheme of things, it's also just maturing you to, okay, this is probably great that we shouldn't kind of support that particular thing that wanted to build or something like that. Right. Yeah.

‍

Christina Cacioppo:

Well, thank you so much today for this conversation on everything from AI to building in Europe to Zambian chickens. Again. I had a lot of fun.

‍

Steffen Tjerrild:

Yeah, likewise Christina. Thanks for having me.

‍

Christina Cacioppo:

Thank you. We hope you found this conversation as enlightening as we did. If you're interested for more about Frameworks for Growth or Vanta, you can find us at vanta.com and Vanta across all social channels. Thanks so much.

‍