#263 Jarek Kutylowski: How DeepL Is Using AI to Break Language Barriers

What if language was no longer a barrier to global business? In this episode, DeepL CEO Jarek Kutylowski joins Craig Smith to unpack how AI-powered translation is reshaping the way companies communicate across borders. From real-time speech tools to enterprise-grade language workflows, Jarek shares DeepL's journey, the tech behind their LLMs, and why accuracy, nuance, and localization matter more than ever. If you're building for a global audience, this conversation is a must-listen. About Jarek: Dr. Jaroslaw "Jarek" Kutylowski is the founder and CEO of DeepL, the Cologne-based Language AI company transforming global communication and breaking down language barriers for businesses worldwide. Born in Poland and having spent a large part of his life in Germany, Jarek brings an international perspective and a lifelong passion for technology - a developer at heart, he began coding at just 10 years old, building tools he found useful in daily life. He holds a PhD in Computer Science with a focus on mathematics and has held roles at several tech companies prior to founding DeepL. Under his leadership, DeepL has grown rapidly, expanding its Language AI platform to offer highly accurate and nuanced human-like translation in both written and spoken formats, as well as a contextual AI writing assistant. Today, over 200,000 businesses and governments - and millions of individuals across 228 global markets - trust DeepL for secure, seamless and effective communication. Stay Updated: Craig Smith on X: https://x.com/craigss Eye on A. I. on X: https://x.com/EyeOn_AI (00:00) Introduction to DeepL and Jarek's Vision (05:09) DeepL's Technology Evolution (08:17) The DeepL Platform: Products, APIs, and Enterprise Use Cases (12:39) Why DeepL Outperforms Consumer Tools (15:14) How Enterprises Use DeepL (17:05) DeepL's Translation Accuracy (21:20) Breaking Language Barriers in Global Workforces (23:46) A Future Without Language Barriers (27:19) Expanding Language Coverage and Industry Verticals (31:44) DeepL's Role in High-Compliance Sectors (34:04) Tackling Latency and Real-Time Use Cases (38:38) DeepL as the Translation Engine Behind Other Products (42:40) Partnering with News Media and Government (45:40) Pricing Models and Accessibility (46:24) The Human Impact of AI-Powered Language Tools (49:22) What's Next for DeepL and Global Communication

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  The difference that DeepL is making comparing to maybe some of those more consumer oriented solutions on the market is the accuracy part, is the integrations into workflows, is the ability even for a company to tweak their language of the translation towards what they need. I think we're moving into this direction where language is not going to be an issue at all. And I wouldn't want to forget about the kind of human aspect of how much language is embedded into who we are and how we can exchange with others. Having access to technology which makes it easier for people who are not so well versed in a foreign language, that helps a lot them to be competitive in the work market,

• 0:45Speaker 00

  but also them being able to express themselves freely and preserve the language. Hi, I'm Jarek Kutylowski. I'm a founder and CEO of DeepL. I come from a computer science background. I am super excited about what technology can actually do to make our lives better and how that can have an impact. And out of that passion, I founded DeepL and I'm kind of right now looking into how AI can help businesses become more international, can help in all of those language kind of issues and challenges that we have every day. And yeah, very much looking forward to the conversation.

• 1:29Craig S. Smith?

  Yeah, I was reading you, you, DeepL, the tech started at a company called Lingue, is that right?

• 1:40Speaker 00

  Lingue, it's called Lingue. Yeah, DeepL is a spin off out of that, out of the company basically. And we've been looking into like the translation space or the language space already earlier in that business. And then at some point in time, 2017, we've noticed, hey, there is this new amazing technology, neural networks, the foundation of AI as we know it nowadays. And you can actually do more about that and there is an opportunity for us to go into that space.

• 2:12Craig S. Smith?

  Yeah, and so describe DeepL's product and tech. I'm curious, I mean, we spoke briefly earlier about that space because ever since transformers hit the scene,

• 2:32Craig S. Smith?

  there have been all kinds of companies popping up with translation solutions. So tell me about DeepL.

• 2:43Craig S. Smith?

  What kinds of applications you're you're applied to? And and what differentiates DeepL from from the crowd?

• 2:54Speaker 00

  Yeah, yeah. I mean, it's been quite a journey, of course, like the company already exists for eight years. And the product is like something totally different nowadays than it was than it was back then in 2017. We've been around quite a little bit earlier, even the transformers reached the market. We've been kind of pioneering the way or like how you designed neural networks for translation specifically and maybe language in a more general sense. And and for that purpose, we've been like doing this really cutting edge AI frontier research in terms of like, what are the model architectures? What really works better? What is a better way of training those models? What is the data that you should feed those models?

• 3:40Speaker 00

  To always achieve the best quality and accuracy in translation. But also paired with like you, we really want to make sure that translation has this nuance, that it feels natural, it feels feels native. And and sometimes that is clashing with the accuracy aspect also. You're translating marketing materials in a different way than than than technical documents or legal documents, maybe. And we've been making sure that our AI models actually can capture all of all of that. With really the aim on looking at what businesses need. While we are a PLG company.

• 4:21Speaker 00

  And therefore have grown through this very strong user base uh that we've that we've been able to build. Uh nevertheless, our goal has been always to make sure that we can provide language solutions to businesses. In all of those areas where regularly accuracy matters, where uh embedding into the workflows of your of your employees uh matter. Think both internal use cases where it's about the communication between different offices in different countries. Uh or even external when you're talking to customers that are outside in in a in a in a different country where you maybe don't have field operations in.

• 4:59Craig S. Smith?

  Yeah, and this is uh both uh text to text uh and speech to text. Does it also do speech to speech?

• 5:09Speaker 00

  Yeah, our our first uh approach to that was really text to text. It has evolved then uh towards uh towards documents. Uh I think that's something that's also incredibly important for our customers, like a lot of the the kind of written content is actually out there in the form of documents and and and it's it's it just makes a lot of sense to approach document translation in a slightly different way. And uh since last year, we also have our speech to text uh translation out in the market. Uh we're looking into the speech to speech uh space. And we still feel it's a little bit early for that from a technological perspective that actually the

• 5:49Speaker 00

  um speech to text, kind of live captioning during uh during a meeting. Um gives you the better user experience and actually is allowing you to better follow on what's happening in a in a in a meeting. Uh but definitely something that we're looking into and kind of waiting on where where is the when is the technology gonna be actually at a point where um it brings a lot of value to the kind of typical business use case.

• 6:16Craig S. Smith?

  Yeah, uh yeah, and the the speech to speech, particularly live speech to speech, the the issue is depending on the language pair, you have to wait until the end of a sentence before you can translate it, and that creates a latency issue. Uh is that right? Like?

• 6:37Craig S. Smith?

  Chinese to German or something where the in German the verbs at the end and Chinese it's at the beginning and.

• 6:44Speaker 00

  Yeah, that that that is a big problem. And I think in those languages, it is the the biggest one. Also, in other language pairs, like some of those problems come up uh come up too. So there is like a little bit of an inherent latency embedded in the problem in some ways.

• 7:01Craig S. Smith?

  Yeah.

• 7:02Speaker 00

  Unless unless we are able to predict what you're gonna uh what you're gonna say. And uh that is also a fascinating field of research, because quite often, the models can do quite a good prediction on what the end of the sentence is gonna be. Nevertheless, there's gonna be mistakes. There's gonna be kind of

• 7:21Craig S. Smith?

  Sure.

• 7:22Speaker 00

  curveballs coming up and uh and then they will go uh they will have to deal with that. But there's there's a few approaches of of of which we are thinking about how you can actually uh do that better with lower latency, uh because latency is is kind of the the biggest problem in this whole uh this whole space.

• 7:41Craig S. Smith?

  Yeah, uh but as long as you're doing uh the output in text, and it's not, you're not expecting live translation, it gives the model time uh to to to look at the entire sentence and and that are you are you using um uh pure transformer models? I've been talking to people about Mamba models uh that have a a longer uh memory in effect. Um what what what's the architecture that you guys are using?

• 8:17Speaker 00

  Yeah, we're we're actually not disclosing the architectures of our models. That's like kind of a little bit of our secret sauce. Uh we've we've transitioned to to L L M's, but there is our purpose-built L L M's that are built for for translation uh over the course of of last year, uh kind of going with the with the with the flow of the advances in uh in the technology. And that has given us once again uh a bigger boost uh there. Uh there's also models involved for speech recognition, of course, when we're thinking about uh speech to text there.

• 8:49Craig S. Smith?

  Yeah.

• 8:49Speaker 00

  And and we're trying to make sure that our translation models are also really uh tailored to uh how you need to translate spoken language, just because it's a little bit different than what you'd usually write. And the models need to be able to cope with that, too.

• 9:06Craig S. Smith?

  Yeah, so the product, how how do you sell the product? I mean, is it an API? Is it

• 9:17Craig S. Smith?

  a usage based model? Or is this

• 9:22Craig S. Smith?

  you know, piece of software that that people embed in their own systems?

• 9:29Speaker 00

  Yeah, it's it's actually like the part where I said like it has evolved over the last years. It has evolved into a complete platform there. And and the way that we're thinking about that is really enabling a whole enterprise to solve all of their translation and language challenges that they that they might have. And that starts with like a seat based subscription for all of those of your employees that really need access to this instant translation, like all of those who are exposed in some ways to to an international context, who don't want to go to an agency, who don't want to

• 10:09Speaker 00

  work maybe with an internal localization team, that's kind of, there's always like a turnaround time for that. If you want to have the translation immediately, then then you can be equipped with that. And it starts with that, giving all of the employees access to to the platform. But then when you think about this further, there is some workflows that an enterprise is going to be running that can benefit from a deeper embedding of the technology into into those flows. A typical one would be like a customer support use case where there is inquiries from customers coming in from different parts of the world. And you always want to route those to the agent responding to those in the

• 10:55Speaker 00

  in the relevant language and therefore kind of being able to to serve your customers better. So so from that perspective, we think about that in a kind of holistically way. There's different ways to to to approach the technology itself from an end user based application over to our DeepL Right product, which which helps you express yourself in your native language with DeepL Voice, speech to text, and then obviously, all of those services can be also accessed via the API to embed into into your workflows.

• 11:31Craig S. Smith?

  Yeah, and the

• 11:35Craig S. Smith?

  Give me, what's the most common use case that you're serving? And frankly, what, you know, I mean, this machine translation has revolutionized anyone who's working across languages. I spent a lot of my time in China. I can read basic Chinese, but I, you know, it's, it's slow and painful.

• 12:06Craig S. Smith?

  But now I read Chinese websites with no problem because built into

• 12:14Craig S. Smith?

  my browser is, is a translation function. I, I think it's from Google. I don't know. And it does a pretty good job. I mean, how do you, that's what I was saying. How do you?

• 12:30Craig S. Smith?

  Why would companies use DeepL as opposed to just using the Google plugin, for example?

• 12:37Speaker 00

  I mean, I, I can only confirm that kind of this development of technology has been amazing. Like we, like 10 years ago, we wouldn't have thought of all of that being really possible. And like one of the kind of latest points for me to really, really realize was with me sitting in a customer conversation in, in Japan where until now, we actually either had an interpreter in the room or I had members of my sales team kind of trying to help me and, and, and interpret while we've been going. But it's always cumbersome. That really doesn't

• 13:17Speaker 00

  work that nicely and is just a bad experience. And, and right now with technology, I had access to like the full breadth of the conversation. I could express myself freely, knowing that the technology is going to be able to pick up on the nuances in, in my English and, and translate them correctly into, into Japanese. That's, that's, that's been pretty amazing. And, and I think kind of the difference that DeepL is making comparing To maybe some of those more consumer oriented solutions on on the market is is the accuracy part is the integrations into into workflows. Is the ability even for a company to tweak their their language of the translation towards what they need.

• 14:01Speaker 00

  Uh whether that's compliance rules or whether that's um certain aspects of how the brand voice looks like. Uh but then going like really also deeply into terminology. And if you think once again about legal text, about uh technical text. Uh that matters quite a lot because it reduces the amount of work that you have correcting the text. And uh that plays a big role for our customers in terms of ROI. Uh because kind of any additional editing that must be done on top of the AI's output. Uh that's incredibly expensive comparing to

• 14:41Craig S. Smith?

  Yeah.

• 14:42Speaker 00

  to to the technology costs.

• 14:44Craig S. Smith?

  What's a typical use case of uh if a company has uh an office in Japan?

• 14:53Speaker 00

  Hmm.

• 14:54Craig S. Smith?

  Uh does it can will, for example, will English language emails automatically be translated into Japanese? Or if you're uh uh sharing a document, is the document uh automatically translated? I mean, give me a couple of uh use cases.

• 15:14Speaker 00

  Yeah, yeah, I mean, uh I I think I think the example is is is great. And and I think like one of my uh one of my favorite customers is is an automotive automotive company, which has all of the R and D really in Japan. Uh whereas large parts of the customer base in the markets are in the US, in Europe. And uh there is really this language barrier between R and D and between the fields and obviously like.

• 15:42Speaker 00

  They need to talk to each other. Um, and in those cases, it really depends also on the customer. There's those who have translation really embedded into their own systems in the company. So that, as you say, like every email that gets sent out and it's a different language, it gets automatically translated. Documents that are posted into uh into a system, uh be it like a contract management system, for example, are being automatically translated. There's tons of plugins that work with D Bell in those systems and integrations uh that do that automatically. Uh, but sometimes it is really the kind of the user who uh either uses our browser

• 16:23Speaker 00

  extensions, our desktop apps, or just goes purely to the website and uh takes the the content of the current document that they've received, uh translates it and uh uses it for its own consumption or even publishing it within the company.

• 16:41Craig S. Smith?

  Yeah, uh and in that case, uh where it's it's documents, uh like as you said, technical documents and that sort of thing, uh do you build uh an ontology for each customer uh between uh the language pairs or how how does that work so that you can ensure accuracy?

• 17:05Speaker 00

  Yeah, yeah, there's there's terminology basis that customers can embed into D Bell or that you can even ask D Bell to kind of infer from your existing uh translations. As a company, you usually come in with a history of maybe um a massive set of uh of of documents haven't been already translated in the past. And and you can ask the AI that that we provide to to give you like to to to extract the terminology out of those uh documents, like what is what is special basically about your language comparing to maybe the average uh that you would want to check and then embed into further translations that are being uh produced by D Bell.

• 17:48Speaker 00

  Uh, but also you can comment with your with your pre uh. Pre-made, pre-categorized glossaries terminology basis, just plug them in into the AI. It's gonna know those words, it's gonna be able to

• 18:04Speaker 00

  properly use them in all of those sentences. It's gonna be able to even sometimes figure out when there's false friends in those translations, when in certain areas you actually don't want to use the terminology that is embedded into the glossary, because it doesn't make any sense really.

• 18:25Speaker 00

  And then deviate from what it was supposed to do in that situation. So you have a lot of control, but at the same time, the AI remains incredibly smart about the language being used. And make sure that whatever you're translating maintains both this DeepL typical accuracy, but also this nuance and the look and feel of somebody native has been writing that.

• 18:52Craig S. Smith?

  Yeah, you were saying it's a platform, meaning

• 19:00Craig S. Smith?

  I mean, but you also mentioned a browser extension.

• 19:08Craig S. Smith?

  So how does a company work with with DeepL? Do they work through a DeepL website? Or yeah, just describe.

• 19:22Speaker 00

  I think I think translation is such a broad use case that and therefore there's also different ways of applying that and different ways of applying it to different workflows. And we, as a provider of this kind of complete solution package of this platform, are trying to accommodate to also what our customers really want there. So which is why I mentioned like there's the browser extension for those who who really wants to work in their browser, but you don't have to do that, you can go to the website. So there's different means and that really depends on the company, how their technological stack looks like, how how much how much they want to deeply embed translation into some of those workflows,

• 20:05Speaker 00

  or whether it's just a casual kind of long tail workflow in that particular case.

• 20:12Craig S. Smith?

  Do you see wide adoption of this technology among multinationals that have offices? I mean, you know, as I said, I worked for a long time in China, I've worked in Europe. And generally multinational companies have senior staff that speak English. And even mid-tier staff speak English. So, English had become the langua franc of business.

• 20:51Craig S. Smith?

  With this technology, do you see, first of all, do you see multinationals adopting it? And is it expanding their footprint, do you think? Or do you think it's leading to less emphasis on the use of English by non-native English speakers?

• 21:20Speaker 00

  Yeah, I mean, like, multinationals are a big part of our customer base. And I think sometimes, like, we expect everyone to accommodate that the company's lived language is English. But we sometimes underestimate how big a problem or a bit of a kind of maybe even efficiency challenge that is for some of the employees that are out there, sitting in a non-English speaking country. And for them, quite often, it is just much more efficient, even if they can speak the language, to write it in their native language, then translate it.

• 22:03Speaker 00

  Or it just kind of gives them this confidence in being able to express themselves in a way that is much more professional, that is on a comparable level as their peers who are maybe coming from the US or from UK or another English-speaking. A country. So there is there's a big emotional component to that quite often. But there is also just this pure proactivity efficiency aspect that is that is still really, really important. And and we're seeing that across multinationals and like in in general international companies.

• 22:43Speaker 00

  That whenever there's a bit of spotlight on getting that diverse employee base engaged in conversations. Making sure that they participate. Then the technological solution to that becomes quite often the the the choice very, very quickly.

• 23:04Craig S. Smith?

  Yeah, do you you've been in this um end of the the AI world for a while. What fascinates me about it is the gradual breakdown of language barriers between cultures and the increased understanding between cultures. Is likely to come from that.

• 23:28Craig S. Smith?

  How do you view that? I mean, do you do you think that that it accelerates or expands business communication and you see a day when language will no longer be an issue at all?

• 23:46Speaker 00

  I I I think I think we're moving into this direction where language is not gonna be an an issue at all. And and and finally, it, I think, solutions like Deep Bell's really help even maintaining those local languages. Because, yes, you have to speak English in your in your business world, but maybe you don't have to put so much emphasis on that. And and and you can live in your local language much, much better. Also, I think over the last years, customers' expectations when it comes to the language in which they're being handled and and and spoken communicated to have grown.

• 24:25Speaker 00

  So all those multinational companies that we've been just talking about, they cannot afford anymore to just offer customer service and customer. Support in English, they will have to accommodate to the local markets. This is just how the demand is growing. People know that those solutions are out there and it's possible. And the competitors are just frankly doing that. So, so I think that technology is actually contributing very much to, to the local strengths in, in, in different languages. And, and I really like that aspect. I think that, that brings in this, this diversity there. But at the same time, it makes our life so much easier when it comes to, to, to international business and, and global trade.

• 25:10Speaker 00

  And we've seen a very strong correlation in our customer base, depending on how strong trade relationships are between countries. And also then the demand on language translation as a, like for, for a country as a, as a whole.

• 25:28Craig S. Smith?

  Yeah, well, how many language pairs do you cover?

• 25:33Speaker 00

  We cover over thirty language, languages right now. So kind of the, the pairs is gonna, it's gonna be like the square product of those. So it's a, it's, it's, it's a large space. We, we are mainly focusing really on quality and specifically for those languages that are the most important in the, in the, in the business area. And we're also the amount of data available and the focus that we can put on those languages allows us to achieve this kind of typical deep quality that our customers and our users know and expect from, from the service.

• 26:10Craig S. Smith?

  Yeah, the, what, what's the, the smallest language in terms of the population, speaking population that you have, or one of the more?

• 26:27Craig S. Smith?

  Exotic, I don't mean exotic because I hate that word, actually.

• 26:33Speaker 00

  That's actually, yeah, that's, that's a good question. I don't have an answer for you on that. Like, it's gonna be like one of the smaller European languages, for sure. But I haven't like really checked on which one of those is, is really the, the smallest one in terms of number of speakers or the population base.

• 26:51Craig S. Smith?

  Yeah, and, and are you growing the number of languages you cover? Or is your focus on, uh, you know, on verticals, on maybe medical uh, or, or, uh,

• 27:09Craig S. Smith?

  you know, scientific, uh, uh, areas where the language becomes more specialized?

• 27:17Speaker 00

  Yeah, uh, actually both. Uh, I mean, the, the language portfolio is, is expanding all of the time and, and we're bringing in new languages, uh, like every year. Uh, this year is gonna be a little bit more about Asian languages. Uh, there's still a huge market. Uh, there, uh, there is, there's more and more collaborations and cooperations that we're seeing within our customer base that span between Asian countries, uh, directly and not only between kind of Europe and the US and, and Asia. Uh, so there's more and more demand for, uh, Japanese to Vietnamese translation, for example. Um, Japanese Korean, uh, translation. So that's, that's something that's, that's kind of coming up as actually the, uh, economic

• 28:02Speaker 00

  uh, ties between, um, Asian countries are also strengthening, um, at least from our perspective over, uh, over the last, uh, over the last years. Um, at the same time, we're looking, uh, also as a company more and more into, into, into different verticals and, and, and the particular language space that's, uh, that's there.

• 28:27Speaker 00

  We can, we can take maybe life sciences as a, as a good example of that with like very specific, specific terminology. There even FDA related requirements on how translation needs to, needs to look like. And, and we're working with our customers to broaden our product also in that direction. And sometimes that's really like hard AI work to make sure that the models are capable of this, this particular language type. And sometimes it's really also industry specific integrations and, and understanding how does translation plug in into, into some of those workflows specific for those areas.

• 29:08Craig S. Smith?

  Yeah, actually that's fascinating. I haven't, I mean, I sort of assume that, that your user base is primarily native English speaking companies trying to reach or, or extend their or make their communications easier with non-English native English speaking companies. But there's a lot of, the world is, doesn't touch English at all. It's between non-English language pairs. What's the, the biggest pair that you see there with the highest demand for?

• 29:49Speaker 00

  I mean, like English German is still our strongest and that's also a kind of, because maybe that's one, like our home, home turf that we're operating in. But that's not only because of that. If you look at the distribution of translations being found on the internet in different language pairs. Then, then English German is actually one of the biggest one, if not the, if not the biggest one. Yeah, yeah, it's actually just because of the kind of economic ties between Germany and and the rest of the world being pretty strong. It's a it's a country that's been exporting for so many years and had this very deep trade relationships with the rest of the world.

• 30:33Speaker 00

  And and most of what was happening externally kind of really goes into or is being translated into into English. Chinese, Japanese are also then always into English are also incredibly strong pairs. And then pretty quickly, we're gonna see like some non-English pairs. Japanese, Korean is one of the important ones. German, French, just because that's like kind of directly on the border. And there's no need for using English as an intermediary language in the conversations between like a German company and their French customer. So you're gonna see a lot of direct German-French material going through DEBEL.

• 31:18Craig S. Smith?

  Yeah, that's interesting because until now, English has been the intermediary language. I mean, a German company speaking to a French company, generally they'll speak English in my experience.

• 31:33Craig S. Smith?

  In

• 31:37Craig S. Smith?

  what what verticals do you see the most demand in? I mean, you mentioned life sciences.

• 31:43Speaker 00

  It's it's it's mostly really those that have the highest dependency on translation quality, but also compliance aspects where where kind of every word matters.

• 31:56Speaker 00

  We're DEBEL is pretty strongly represented within legal firms, within professional services companies.

• 32:05Speaker 00

  Everywhere where there is like a big

• 32:10Speaker 00

  Numbers of documents being exchanged where there is a lot of white collar work, knowledge workers happening and at the same time where where the stakes are high. And and I think legal, life sciences, professional services, but also manufacturing. Because really manufacturing companies very often span different continents. They spend different different languages at that makes them

• 32:41Speaker 00

  relevant for us and the kind of language barrier being important for them.

• 32:45Craig S. Smith?

  Yeah, the do you service the the European Commission or any of the EU bodies?

• 32:56Speaker 00

  And I think not directly to to a large extent. I think certain EU bodies might be using DBEL either through through us or through partners. There but we're working with other governments and have actually government contracts, especially for countries which are multilingual.

• 33:19Speaker 00

  That's that's a big market for us, just given their internal needs for even publishing documentation and materials and the legal ramifications on what they're obliged to do.

• 33:36Craig S. Smith?

  Yeah, yeah, that's why I mentioned the EU. I mean, don't they have to publish every document in, I don't, five languages or something? I can't remember. But yeah, and and I would guess on latency. I mean, you're not doing real time right now, but but you did say instant. So what what kind of latency are we talking about?

• 34:03Speaker 00

  It's like when it comes both to text and and and speech, I I think we're thinking quite often two hundred milliseconds, uh, something, something in that, in that range. Uh, that's, that's not a lot. And I think that really helps you, uh, being able to, to, when it, when it's, when it's about speech to text, to, to follow very well, uh, that conversation. It's not zero, though. Uh, that is, that is still latency and, and we're working very hard to get that down. And, uh, that is both work on the level of making the models faster, smaller, uh, and, and being able to just be more efficient, but also deploying the models as near to the user,

• 34:47Speaker 00

  um, in geographically distributed data centers and, and that, that's still, that still matters.

• 34:54Craig S. Smith?

  Yeah, although two hundred milliseconds is pretty fast still. Uh, the, um, uh, so are, are, are people using DeepL not only for, uh, text to text, uh, in which case latency isn't as, as, uh, critical, but for subtitling, for example, in a Zoom call or, uh, yeah.

• 35:22Speaker 00

  Yeah, with, with, with DeepL Voice, that's exactly the, the case. It's, it's real time, a meeting, uh, subtitling, uh, and, and therefore, like, this is, this is really critical when it comes, when it comes to speed. This is, this is essentially real time, um, in that case. Uh, but also with text to text, latency matters. And, uh, for example, in, in customer support chat, uh, use cases, uh, we are, we are fighting for every one hundred milliseconds there too, uh, because that all adds up for a great customer experience. And, uh, whatever, whatever is the inherent latency of the response of somebody whom you're talking to,

• 36:06Speaker 00

  you don't really want to add up a lot on the translation itself. So minimizing that is an, an important part. And, and actually something that our customers really care about and, and, and are asking, uh, for.

• 36:18Craig S. Smith?

  Yeah, and on that 200 milliseconds.

• 36:26Craig S. Smith?

  I mean, as we said at the beginning, depending on the language pairs, there is kind of a fixed latency because depending on how fast the speaker is speaking, because you need to get to the end of a sentence before you can accurately translate the full sentence. So how do you bring down that latency? What is it through prediction?

• 37:00Speaker 00

  It's a mix. It is, there is, there's a bit of just simply sometimes waiting for the end of the sentence and then translating as quickly as possible. But there is also an aspect of prediction in that. And sometimes the kind of advantage that you have with subtitles is that if you're wrong with your translation, you can still correct yourself. Like you can basically flip the subtitle a bit. And we're trying to minimize that as much as possible because it is a bit confusing if that happens too often. But in rare cases, when this already occurs, when the prediction is not correct,

• 37:45Speaker 00

  you can, you can pull that. And that helps to bring down the overall latency of the, of the whole conversation.

• 37:53Craig S. Smith?

  Yeah. Does DeepL provide its tech as, as a component for other software systems that are then being sold on? Where, I mean, I'm thinking of, you know, I talked to a A guy recently who's developing a live translation plugin for Zoom. And I mean, is it possible that what he's doing is taking DeepL and adding some bells and whistles and rebranding it? I mean, do you do that? Provide that as

• 38:37Craig S. Smith?

  as a part of a tech stack?

• 38:40Speaker 00

  Yeah, totally. I mean, that's that's kind of basically the API play that we have. And and I talked about that being like relevant for enterprises to to put in translation into their internal workflows, but it applies in the same way to to tech companies who just want to have translation as as part of their of their product. And under the hood, a lot of the translation that you're going to be seeing in in some tools is going to be actually provided by by DeepL and in the bank crowd. So we've been at the forefront of this translation space for for quite a while and and have been able to build out those relationships with ISVs with

• 39:20Speaker 00

  with companies that that build great products, but also need a translation component in those. And this ranges really from

• 39:31Speaker 00

  software that is intended for very dedicated, small audiences, like for lawyers, for example, across to general productivity suites, which cater to like basically all of all knowledge workers out there and also need that translation aspect as as part of of what they're what they're doing.

• 39:57Craig S. Smith?

  Yeah, and is is the I would guess, but I want to hear it from you. Is your market expanding quickly? Or is are there so many players in the market that that it's slowing down growth?

• 40:16Speaker 00

  It's, it's, I mean, the, the, the, the big players and the competition is something that we've, uh, that, that we've been born into. Uh, it is, it is basically something that we've, that we've lived from, from the beginning on. Um, and that competition is out there, that's gonna be always out there. Uh, it's, it's part of our DNA to, to be able to cope, uh, with that. Um, I think what is exciting about this market is that the, the possibilities and the abilities of, of the AI, um, are ever increasing. And every year we are basically seeing, uh, new advances. And those advances, like every time we build something new, uh, they really enable new use cases, uh, where machine

• 41:01Speaker 00

  translation can come in. There's, there's use cases where five years ago, you wouldn't ever think about using machine translation for that because you think it's just too critical. Uh, and right now our customers are, are finding out, hey, that's, that's, that's like totally enough. And, uh, we can speed up our development cycles. We can reach new, uh, customer bases. Uh, we can be just so much more, um, efficient, uh, with, with the technology. And, and therefore I think the market is really expanding and also expanding for us pretty, pretty rapidly.

• 41:36Craig S. Smith?

  Yeah. Do you, um, uh, do you work with any news organizations? Uh, and I ask because I'm, you know, I, I was at the New York Times for much of my career. Uh, and I, I actually ran, or established the, uh, Chinese language platforms. Uh, and, you know, we used human translation. We had a team of translators that sat every day. They translated thirty odd articles. Uh in Beijing. Uh and I remember being really eager for machine translation to start working. And I I left uh before that happened.

• 42:21Craig S. Smith?

  But um but it seems that that this would be perfect for news organizations to broaden their their footprint. So or their audience. So are you working with news organizations? And if so, can you talk about any that are using the tech?

• 42:40Speaker 00

  Yeah, yeah, it's it's it is one of our customer bases and many of them I cannot talk about. Um, the one that I can talk about is actually Nikkei in Japan, uh who have been translating all of their articles or like making them available to to a more international um reader uh base uh with with technology. And I always found that fascinating in a way that uh this this kind of ability to read up on what's happening in another country in their own with their own local press that always gives so much better perspective on what's actually

• 43:22Speaker 00

  happening there rather than going to your local newspaper and and and and and getting like a uh a third party view on on on what's happening uh there. And that's that's both kind of enabled by some of those publishers who are who are doing better with D-Bell. Um, and on the other hand, there's obviously also uh D-Bell users who are using our browser extension and and and kind of uh reading up on their own without the the kind of publishers' direct support.

• 43:53Craig S. Smith?

  Yeah, the the news organizations that are using it, are they? Do they just put a link in their, on their website? If you want to read this in English, French, German, or, you know, click here? Or are they publishing entire sites in, in, you know, English? If they're, say, Japanese, to, to give an English language audience access?

• 44:23Speaker 00

  Yeah, the, the answer is a clear both.

• 44:26Speaker 00

  So we've, we've seen, we've seen both. There are, there are publishers who just give, like, offer that as an on-demand functionality for the users. But there's also publishers who just want to, like, run subsites of whatever their offering is and do that in a different language. And sometimes it's, it's like fully automatic translation. Sometimes depending on the visibility of the article and the criticality of that. There's a human aspect in, in that and there's a review going on. We've seen basically all of those and, like, our technology enables all of those use cases. It's up for the publisher, for the customer, really, to decide on how they want to interact with an international audience.

• 45:08Speaker 00

  And that really very much also depends on the publication and, and the country you're, you're in.

• 45:13Craig S. Smith?

  Yeah. And, and is it priced so that? I mean, you said a subscription. Is, is, are there subscription tiers based on? And you also mentioned seats, you know, depending on how many people in an organization. So can you describe the, the sort of pricing matrix a bit?

• 45:39Speaker 00

  Yeah, yeah, basically two models there. I mean, when it comes to equipping your employees with just the kind of easy access to the tool to do their own translations, then it's seat-based, as, as I mentioned. But as soon as it goes into the direction of the API and integrating that into workflows, integrating into other products, then it's going to be the typical API consumption-based, usage-based model where we price per, basically, character translated.

• 46:11Craig S. Smith?

  Yeah. Okay, I'm running out of things to ask. What would you like listeners to know that I haven't talked about?

• 46:24Speaker 00

  Yeah, I would like everyone to be really fascinated by how AI can really revamp that space and how we communicate with each other. I think there's a very pragmatic aspect to that and like, how well is the communication in my company working? How well can I reach new markets? How good am I competing in this global world as a company? But there's also really this emotional aspect of that and we had employees from our customers on stage in our customer events

• 47:06Speaker 00

  who've been really feeling differently about how they can participate in this global work environment since they have access to this technology. And with all of that efficiency, I wouldn't want to forget about the kind of human aspect of how much language is embedded into who we are and how we can exchange with others. So and therefore, like, I'm super excited about what the technology brings in the future and how we're going to be able to grow together even a little bit more.

• 47:40Craig S. Smith?

  Yeah, and you mentioned, I'm not sure this is what you were talking about, but how it protects or preserves

• 47:53Craig S. Smith?

  languages that are otherwise being eroded by English. I mean, I've spent time in the Arctic and there's, you know, Inuit and some fascinating languages up there, but they're shrinking.

• 48:10Craig S. Smith?

  And something like this would allow them to operate on a global scale without losing their native language.

• 48:21Speaker 00

  Definitely, definitely. I think it's gonna be more challenging for those extremely smaller languages, which you're just talking about. But there's even like this mid-level of actually big countries with their local big languages. And I've been born and raised in Poland, so Polish is my native language. And even there, I see like, everybody's learning English. And that's good, it still should, I guess. But having access to technology, which makes it easier for people who are not so well-versed in a foreign language, that helps a lot them to be competitive in the work market.

• 49:04Speaker 00

  But also them being able to express themselves freely and preserve the language.

• 49:09Craig S. Smith?

  Yeah, so looking forward, what does DeepL have in store for us? And how do you see the world evolving with language barriers becoming increasingly porous?

• 49:22Speaker 00

  Yeah, I mean, we've talked quite a bit about speech translation, and I think that's still an incredibly interesting avenue. I think we're a little bit right now at a level where we've been with text translation in maybe 2018 or so. It's still the early days, and I'm excited to see how that really plays out in real-life usage in all of those multinational businesses. I think when it comes to text translation, I'm still really excited about the path of making sure that every company can speak their own language and how customizable AI can become in an easy way. We've spoken a little bit of terminology today, that's kind of the beginnings.

• 50:06Speaker 00

  I think we're gonna be able to, in a much better way, make sure that all of those businesses are unique out there and also unique in the way that they're communicating and that the AI really understands what is being communicated. So I think a lot of advancements that are coming up in this area. And of course, like, quality advancements in all of the languages that we have out there. There's still a long way to go. I think we've come to an excellent translation quality for the major languages that we know. But for smaller language barriers, there's still work to do, and we really should not forget those.

• 50:49Craig S. Smith?

  Yeah, uh, yeah. And, and just on that, on, on the, the quality of translation.

• 50:58Craig S. Smith?

  Is it? Is it that, that, uh, the translation, when you run into quality issues, is it that the translation is understandable, but is clearly not, uh, from a native speaker? Or is it, uh, that there are, uh, mistranslations, uh, that, that, uh, could change the meaning of sentences?

• 51:30Speaker 00

  I like the, the mistranslation case is getting rarer and rarer as the technology develops. Uh, like, obviously, we cannot fully rule that out, as much as we also cannot rule that out with, like, human, uh, translation mistakes just happen, uh, there. Um, I think exactly it was what you're saying. It's like, sometimes it just doesn't sound that great yet.

• 51:54Craig S. Smith?

  Right.

• 51:55Speaker 00

  And, and we gotta get the technology to the level where, uh, it always just feels supernatural and, and, and as if a native speaker would have, would have written it.

• 52:05Craig S. Smith?

  Yeah, um, and, I mean, we'll wrap up, but I'm just, as you're talking, I'm thinking that it would be great to have a site that translates all the, the world's, uh, major media in one place. So, you know, you could pick a topic. Or maybe that exists. Do you know if that exists?

• 52:28Speaker 00

  I, I don't, I don't know if that existing. I don't know if the publishers would be so happy about like an aggregator like that. Uh, I think that's, that's, that's a discussion of a different type there. Um, but I think, I mean, in general, it is, it is really possible. Like, take the belt, take the the belt browser extension, put that into your browser and just go to, to all of those different, um, media places out there in the world and, and, and start browsing their content. That's feasible already now.

• 52:59Craig S. Smith?

  You don't sell the browser extension as a standalone product, do you?

• 53:04Speaker 00

  Uh, there is, there's a free tier there, so you just can download the browser extension and start using that. And depending kind of on your usage patterns and like how much um and and what you need. Uh, there's gonna be, there's, there's gonna be an affordable plan there uh to use that.

• 53:22Craig S. Smith?

  Okay, great. Well, I'm gonna download it as soon as we're done with this call.

• 53:28Speaker 00

  Perfect. Eddie, Eddie can assist if there's, if there's anything on the team.

• 53:32Craig S. Smith?

  Sure.