2946: From Translation to Transformation: How DeepL is Leading the AI Charge

[00:00:01] Have you ever considered how the nuances of language can impact global business success? Well today I've got a very special guest joining me. His name is Steve Rotter. He's the Chief Marketing Officer at a company called DeepL and he's

[00:00:15] going to be showing how their specialized AI translator is revolutionizing communication around the world. And since its inception back in 2017, they have been dedicated to providing high quality translation services while supporting businesses and individuals in overcoming language barriers with

[00:00:34] precision and security. And with a suite of products ranging from web translators to AI writing tools, I want to learn more about how they're prioritizing data protection and continuous innovation. So my guest today will talk about the unique technology behind their success, their stringent data security measures

[00:00:56] and the significant business value of machine translation. And we'll explore these advancements and also discuss what the future holds for AI and language services and how AI driven translation can actually help enhance your business

[00:01:14] operations and even global reach. If you want to find out more buckle up and hold on tight as I beam your ears all the way to the US where my guest has got up

[00:01:23] at 5 a.m. to speak with me today. But enough talk from me, let's get Steve on the show now. So a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do? Sure deal yeah my name is

[00:01:39] Steve Rotter I'm the Chief Marketing Officer here at DeepL and my goal is to just continue to educate the market on the power of AI for things like translation and writing and I'd say it's a fascinating and fun job. It really

[00:01:54] must be and it's an incredibly exciting space right now and what you do at DeepL they're translating whole files, change of documents, language, retaining the original formatting for maximum time saving all that cool stuff but there is

[00:02:06] so much going on in this space so many big companies so I've got to ask yourself you must be feeling it right now with so much attention in this space how do you at DeepL manage to make yourself stand out from the crowd from

[00:02:18] all these tech giants are almost encroaching in this AI translation space of what unit advantages do you try and offer them? Yeah I mean I think it's a great question because anybody who's just paying attention to tech sees this

[00:02:31] every day right they see the big announcements the big partnerships the battles out there but I think there's one critical distinction right out of the gate that is also happening in the market which is this kind of split this

[00:02:45] fork in the road between general-purpose AI and specialized AI there is not a debate whether one will be better or the other but simply just the use cases right and this kind of comes down to like who's using the

[00:02:58] technology for what purpose so if you think about when you say translation maybe some of us might have like an immediate reaction to like a vacation where we had to translate a menu or something that that's fairly low risk

[00:03:08] right but then then you just kind of shift over and think about your own workplace right and think about like the tens of thousands of documents that your that your legal team might be translating or your brochures that your

[00:03:21] marketing organization is translating or your IT organization interacting with people in different countries and translating even internal communications all of those use cases really raise the bar for one thing which is translation quality so now Neil you go back to that kind of first point between general

[00:03:41] purpose and specialized models when we think about that at DeepL we've been focused on language since day one right so the end of this this AI it's specifically tuned for language the nuance of language and when you're

[00:03:54] talking about all those kind of high I guess I'll call that more business critical use cases you got to get it right right and that holds true and that's really why we're able to continue to be perceived as a market

[00:04:08] leader in translation because of that quality right if you don't have that then then everything else kind of falls apart there's a few other things on top of that quality that really start kind of making that stark difference the

[00:04:20] security issues I mean you can't open up the the trade publications without seeing this this ongoing debate about well are you using my data to train your models is it available out there for other people is it gonna show up in somebody else's

[00:04:34] chat GPT prompt I mean there's a lot of things in and around security it really become pretty messy so we always lean toward very strict we don't train on your data we delete immediately after translation solid kind of GDPR sock to

[00:04:48] ISO all this kind of certs that go along with land certification and what's actually really cool is we own our own data center so I've been up in Sweden there is a massive Center that's completely eco-friendly so sort of

[00:05:02] completely green tech 75% powered by a hydro 25% and the construction is made in a way that's pretty like it's all eco-friendly and even the heat from the computers are actually piped into a facility to dry pellet wood pellets for

[00:05:19] heating homes in the area so it's actually quite fascinating but when you think about like all of those three things around like not using the model or not using your data to train the model the security certifications and

[00:05:31] all that and our own kind of data center it actually makes a pretty pretty secure environment for people that really trust and I'm glad you brought that up because we're all seeing all the headlines around AI but if we look

[00:05:42] beyond those headlines from home people listening are thinking about hey what does this mean for my workplace what use cases are there that can impact my work there so with that I suppose just to set the scene for our conversation today can

[00:05:56] you explain the role of machine translation in modern workplaces and how you're helping to maybe enhance productivity and communication for businesses because I think it's great to look beyond those headlines and hey what

[00:06:07] problems are we solving it yeah I mean I think the most obvious one would just be like think of a kind of the standard artifact a document right someone needs it in in Portuguese German and Japanese and they have it in had it written in

[00:06:21] their native language and they just want to translate that into those languages for distribution whether that's internal or external I think that's one very obvious one at that level and in many cases the machine translation is just an

[00:06:34] accelerator right it may be not it might not be a hundred percent translated and it might have a human-in-the-loop kind of support in some cases maybe for less business critical use cases it is a hundred percent machine translated so I

[00:06:48] think that you see that kind of spectrum really of the how machine translation and humans can work together I think what's more fascinating is as the technology has become more powerful including that we like to call it anywhere you work so you

[00:07:07] know there's kind of a standard deep L application where you can kind of go in and use that application specifically but there's also a couple of other interesting use cases right so I'm using my kind of Microsoft Word or my Google

[00:07:22] Suite and I'm authoring something and I'm translating on the fly right so now all of a sudden and this is actually a really interesting story which we heard the other day from from an organization using the technology is the word

[00:07:35] confidence now if you're an engineer and you're you're very comfortable in your own native language but you're not comfortable communicating internally with say your English counterparts or your German counterparts then all of a sudden

[00:07:49] you might kind of appear to be a little bit more reserved or you might not be speaking up but all of a sudden if you're able to kind of translate your thoughts immediately bring that into the language of the receiver you almost

[00:07:59] create this connection and that all of a sudden opens up a business opportunities productivity opportunities for people that really weren't there before so there's a couple of I guess I'll say the obvious use cases but then some of the

[00:08:11] more nuanced ones that come with that translation being available anywhere you work so last example of that is actually applications that are already out there so lots of organizations have dozens 50 60 70 enterprise applications running with our technology there's a there's an API to kind of connect those systems to

[00:08:30] this translation capability so you could extend imagine your customer service and support chat capabilities to the world right you could extend your your external facing ecommerce site to the world right so there's a lot of capabilities that when you start communicating with people in their own

[00:08:47] language opens up a lot of doors for companies to really grow and if we were to look back in time five to ten years ago the IT tech projects etc always had a bit of a reputation for coming in over budget and beyond their expected days oh

[00:09:03] really I guess never heard that but I'd like to think that those days have changed a little bit and every tech project now is it's push for things like ROI what business value is it offering especially when businesses are

[00:09:16] challenged we're doing more with the same or even more with less so on that side of things what measurable business impacts and ROI our company's experience by integrating AI translation services like deepal into their

[00:09:29] operations any value adds that you can share here yeah there there are lots of them I mean I'll go back to your first question though Neil which is this kind of this this massive growth of some of these very popular foundation models

[00:09:44] whether it's open AI or athropic or so forth there you just have these these pressure in the marketplace and I'm sure you hear this from the IT organizations all the time they're feeling pressure from their boss they've

[00:09:54] got to do something we've got to do something around AI back to a recent recent research report on this that shows exit like the budgets even are flowing from kind of though out of the line of business into corporate IT just

[00:10:08] because they have that pressure to like hey you guys do something the problem that they're now facing is okay now what I've got this I've got to do something I've got my boss yelling at me to do more with AI but I've gotta find

[00:10:18] something where I don't just like have this 18 month big bang project that doesn't yield any results because the world's gonna change at that time anyway so we think about like language it's a really just perfect use case for getting

[00:10:34] started this is what we tell in our organizations all the time that we work with is like look just get started because this is a proven use case where you can add value in days and weeks so it really comes down to this kind of

[00:10:47] time to value equation now back to the ROI I mean you look at the business results I mean you're talking about like a 90% reduction in time so if you got Joe do a big manual translations and now he's gonna do this with some kind of

[00:11:01] machine translation technology well you've just freed up a lot of his time right and again when we think about the impact of that on the workplace manual translations is not something that is it's you know it's creating opportunity

[00:11:15] but it's slowing organizations down so if you do that a lot faster think about a typical boss yelling at his team like hey we've got to get into this new market faster or we've got to launch this new product by Thursday or we've

[00:11:28] got to get these this new product design out to the to the manufacturers by you know the end of the month like all of that time pressure put that and you kind of make it easier because of this machine translation capability so time

[00:11:41] to value really really is kind of the big driver for adoption but if you look at the ROI of that it's recent report we did with Forrester they quantified this for us it's like a 340% ROI right so some of these numbers are pretty staggering

[00:11:56] but it doesn't take long to figure out like oh yeah I can see how that would be the case because if you're taking 90% of their time out of a business process it pays for itself pretty quickly well that's a phenomenal start and if we dig a

[00:12:09] little bit deeper on that how do your tech and translations tools how do they better support things like global growth and internalization for businesses that are also looking to expand their reach by leveraging this technology yeah I mean

[00:12:23] it immediately I mean imagine imagine you're a medium-sized business and maybe your core markets are like a couple of countries in Europe and you're like hey I'm really doing great in Germany and I'm doing great in in France but I

[00:12:36] really I really want to reach kind of southern European markets or the Nordics or even out I want to expand into Asia but just a thought of that for a business can be a little bit daunting because you start thinking about all the things

[00:12:47] that you need to do well my see my website needs to be a bad language and now I need to have my customer support documentations it took me nine months just to build them in English now I got to do it all of those things become

[00:12:59] barriers to getting stuff done with machine translation with deep you take all that friction out of the process so now all of a sudden the opportunity to grow globally for organizations and you've reduced that those barriers to

[00:13:15] entry significantly right so that's one of the biggest areas that we see is just you kind of take those those big challenges because a lot of times a product is something that could work well in any particular geography it's

[00:13:27] really just those other barriers that make it difficult for organizations to expand now the other kind of the flip side of that is that's kind of the external facing how do you grow and globalize and your customer base the

[00:13:39] other is just globalization of within employees one of the big use cases we see is actually internal communications right so all of a sudden I just can't like pop up into a completely different country and start selling my product I

[00:13:52] need people there well how am I going to communicate with them do I have expectations that they all speak perfect English or French or whatever my native languages yeah probably not right that's gonna be another barrier so that's one

[00:14:05] of the things we see is just the ability to communicate internally we had interesting conversation the other day with one of our customers and they're like yeah we've got people sitting on our internal meetings completely checked

[00:14:16] out because they don't really understand what's going on well imagine all of a sudden if you're able to kind of make that communication cleaner and more appropriate those employees feel like there you care about them they're like

[00:14:29] oh wow this is just a bit the time to actually speak in my own native language they want me to know what this new strategy is they want me to know

[00:14:36] what this is so yeah a lot of a lot of ways that this is opening up doors for globalization and of course it's almost law that 15 minutes into any business conversation around AI we have to talk about compliance and security measures

[00:14:51] etc so what compliance and security measures do you implement to protect things like sensitive data sensitive corporate data during translation activities and and why do you think these factors are crucial for companies are is fairly obvious answer of course but it must be a question you get asked

[00:15:08] a lot yeah I mean it's fascinating right because the world of and especially the the larger AI companies that probably our best advertising for security right because you look at all of these lawsuits and all of these kind of things going

[00:15:22] on in the market about your training on my data and all that as I mentioned earlier we don't really do any of that right so we're deleting the data on translation where we're not training our models based on customer data so all of

[00:15:33] that gives them a sense of trust and security in the process but there's one other element that's actually quite fascinating that we we encounter with some of our very I guess I'll call it the business critical use cases is

[00:15:46] imagine right you're gonna do let's say you're you're you're leading the merger between two important companies right there you know the kind of Wall Street Journal Financial Times level news now you've got to translate that communication that press release those announcements in multiple languages but

[00:16:06] you can't send that out to five different organizations to do that translation because of the confidential nature of that announced so now all of a sudden the ability to translate that using a machine that's completely secure on I guess on you're not having like five different agencies and

[00:16:24] different people touching that sensitive information you've completely eliminated your ability for that information to leak out so we see this use case actually been quite often where there's highly sensitive information maybe it's a financial transaction between companies maybe it's their earnings report that's

[00:16:43] going out and think about like how that if that were leaked to organizations in advance that could cause stock price changes new product design so you've got your your design team maybe working in one part of the world they're working

[00:16:56] with a manufacturer in another part of the world translating that into multiple languages could cause leaks within security says lots of kind of I guess I'll call it sensitive information use cases that we're starting to see as well and

[00:17:09] of course this year alone we're seeing so many advances in AI whether it would be around text audio and even video now so how do you see your business model evolving are there any potential changes in the road ahead I appreciate you might

[00:17:23] not be able to share too much let me just pull up a visual of the roadmap no I look at I love this this market because it's moving incredibly fast I just joined the D-Belt earlier this year to be honest and even in that time

[00:17:35] there's been so many new innovations new languages the companies rolled out new markets we're at 60 markets now and one of the things that's been a recent innovation and I'll share just in the last couple of weeks we launched a new

[00:17:49] completely new product focused on writing so it is that is plenty of I guess I'll call it generic or general purpose AI tools if I want to write a bedtime story about a robot but if I want to take my original thoughts in my

[00:18:04] language and then translate them and then tune that translation so that it's really kind of captures the essence of my language there's nothing else on the market do that and that's one of the things we're seeing is quite fascinating

[00:18:17] from a use case perspective now we'll a sneak preview there's something coming up in a couple weeks that we're gonna also be sharing and then in the fall maybe something even a little bit more innovative so yeah we're we've got a

[00:18:28] roadmap here that is quite exciting but for us the core of it is doesn't deviate too much from this world of language which is again becoming one of the primary use cases for AI in the world. Wow a few teasers there I think you've

[00:18:42] just cemented your your position for a return interview in the fall that's for sure from your perspective though what would you say the key elements required to build a robust AI competitor in this translation industry how have you

[00:18:55] approached the challenge like this? I mean look the barriers to entry are quite high for any AI company right I mean you obviously a technology just to fuel the computing I mean right now you know that supercomputer that we're

[00:19:09] running on it's the 34th fastest supercomputer in the world like you can't just like spin up like all of the Nvidia tech overnight and make that happen so I think the barriers to entry are really high just from a pure

[00:19:22] foundational tech perspective but beyond that even like even bigger than that it's a matter of focus right I mean since day one the company has been focused on language and the precision and nuance of language so you're building

[00:19:35] your the neural networks the large language model kind of technologies in a way that your goal is perfect pixel perfect translation between one language and another and if that's your mission in life you're building things

[00:19:50] in a certain way I mean even to the point where with all of the research teams that we have they're obsessed about language quality so we'll do we do all of the kind of technical learning that you'd expect from an AI technology

[00:20:04] but we're human validating that and human testing how well our model performs all the time so we've got hundreds of human translators they're not involved in the translation real-time they're not involved in like a typical transaction

[00:20:17] but they're validating the model right it's like hey did we get this from from from German to Portuguese really well and we're constantly tweaking and tuning that and ensuring that the quality is there so that's the North Star for us is

[00:20:30] quality and focus so that becomes a huge barrier to entry to any other organization just coming in and saying how we want to do this too. And as someone that is based in the US you I think you got quite a unique vantage point on this

[00:20:43] subject so how do you see European tech particularly in the AI and translation sectors how do you see that European tech competing on a global stage and are there any unique strengths that it brings to the international market as

[00:20:57] someone based in the US? I mean I think by definition if you're in Europe you're so comfortable with the globalization of businesses and the language challenges and the cultural differences and even just the communication I mean I always just as a fun exercise or sitting around the

[00:21:16] dinner table with colleagues it's like how many languages do you speak and it's like oh I only speak four something like that like okay well only four is not that much if you're in the U.S. and maybe two is a stretch but that

[00:21:28] perspective I think is just simply that the DNA for many European companies is multi-language and multi-english and I think that starting point gives quite frankly it's a unique competitive advantage for a company in the US that

[00:21:43] says oh well we built this really cool tech what do we want to point it at? Oh I know there's lots of languages in the world why don't we do that like someone sitting in Silicon Valley might be scratching their head saying oh there

[00:21:55] really are other languages yeah there might be a few let's see if we could point our AI at that. That's not the case in Europe you kind of like born and bred with this kind of global mindset and multicultural multi-language and I

[00:22:08] think you have a in that in that market you have a much stronger appreciation for quality or you have a much stronger appreciation for the fact that like if I get this wrong my customers aren't gonna receive it the same way or if I get this

[00:22:23] language wrong I might actually even distance myself from a potential customer or business partner there's a lot of impact down there where again doing it from any other part of the world you might not have that same

[00:22:36] appreciation. Yeah I was at a tech conference in Helsinki recently I heard somebody say a great line on stage it said in Europe we can't innovate as well as the US so so we regulate. Yeah that might be a little bit of truth in that

[00:22:51] but I also think that that innovation is gonna come in very different areas and I think again going back to what we kind of chatted about at the onset you know that this did fork in the road around kind of general-purpose AI and very

[00:23:06] function-specific specialized AI I think that's gonna be a big thing in the years to come and I think you're gonna find organizations all around the world that have a unique either perspective or regional kind of expertise leveraging AI

[00:23:21] in a way that is something that regardless of the big tech companies and how no matter how many billions and trillions they throw at infrastructure they're still not gonna be able to compete because they don't have that value and that we found that ultimately it's that business value

[00:23:35] that's driving adoption. I saw something listening to something yesterday about this where you had these kind of AI tourists if you will where yeah though and the concept was quite simple it's like okay hang out because I can make a

[00:23:49] cool avatar but I'm not gonna come back and use it every month I'm just gonna literally like come in and then pop out and ultimately that's neat it's kind of nice to see yourself represented as a squirrel or something like that

[00:24:01] but then all of a sudden you're like okay am I really gonna come back and change my avatar every 30 days and the answer is quite honestly no right but then you kind of contrast that with like this idea of ongoing business value

[00:24:12] well certainly translations one of those things and it's like it's gonna stick you're gonna you're not gonna just use it once you're gonna say oh wow this really made my life easier I have lots of things I need to translate in my

[00:24:22] business in my business role I'm gonna keep using it. 100% with you and I think something else that every listener is feeling right now is that real pressure on us all to be in a state of continuous learning so as someone right in the

[00:24:37] heart of this space and having to evolve and adapt in almost real time to the trends around you where or how do you self-educate to keep up to speed with this constantly moving pace of technology? I could recommend a great

[00:24:50] podcast so I mean I think one of the things I encourage everybody and my team and even even personally is you know this is almost like your diet or your exercise regimen you're not gonna do it once a month and expect

[00:25:09] results so I mean my only advice is find a couple of really good trusted sources and make that like a daily part of your year or at least every other day and make it something where you're immersing yourself because it is changing

[00:25:22] very quickly right new research coming out new models new technologies and in ultimately I think for all of us in the marketplace not just been tech just general this is a tipping point I mean this is a shift I mean they say well the

[00:25:37] next industrial revolution or something like that some of those might be hyperbole but some of it might be very much we're looking at a very different world and how people interact with technology and and being in front of

[00:25:50] that I think it is a good opportunity for just about anybody in the marketplace. Well I love chatting with you today and for anyone listening just wanting to dig a little bit deeper on exactly what you do at DPL explore some of the topics

[00:26:03] we discussed today connect with you or your team or ask a question well how would you like to point everyone listening? Yeah I mean connect with me on Netflix like LinkedIn would be great ask questions and send a note that

[00:26:15] would be fantastic I'd love to kind of point in the right resources certainly if you just go to DPL.com you can just actually try it out but you know we we've kind of have that I guess I'll call it that minimalist interface in

[00:26:27] many ways where it's like okay I just want to get to the product and try it it's a free free all the time kind of a translation and if you find that it's helping your business then then of course there's there's paid versions and

[00:26:38] trial versions and things like that but you just want to go try it out and say oh wow this is really amazing how well it does that's one of the things we hear time I mean I just share these stories because I think it's to me it's

[00:26:48] ultimately why we're here but we did a we did a product launch just a couple of weeks ago and we went to we went to Japan we went to Seoul and I did an interview you sit down with a journalist you don't really know what's gonna

[00:27:00] happen in that interview and the very first thing that journalist said was I want to thank you he was thanking our CEO and we're like oh okay why is that and apparently as he told that story it was that a couple of years ago this big

[00:27:15] story broke the fortune 500 CEO was part of this story and it was a big announcement for them and all his colleagues used Google Translate and they got the story wrong and he used deep L he got it right he said he goes I

[00:27:29] want to thank you because this was kind of a career-changing moment for him in that like he he got the story right as a journalist you don't want to mess those things up and we hear those kind of comments all the time so for us that's

[00:27:42] kind of what keeps us going it's just like that the the positive feedback and that the customers that are just saying look this is really making an impact for me thank you keep doing what you're doing. Wow what a great story and so much

[00:27:55] I loved about chatting with you today especially going beyond the headlines of AI learning more about the role of machine translation what it plays in the workplace also the business impact of AI how can the benefits of AI be

[00:28:07] quantified what ROI are companies already seeing today that today that is the magical side of this for me and thank you for shining a light on it today. Thank you very much Neil it's been great chatting and yeah look forward to

[00:28:19] hearing more from you. So how can specialized AI redefine your approach to language and communication as we heard from Steve today I think their commitment to high quality translation and robust data security is helping to transform the way businesses interact globally and by offering precise translations and

[00:28:41] ensuring data privacy they're not only enhancing productivity but also helping foster inclusivity and global expansion so huge thanks to Steve for an insightful discussion today I think the future of AI and language services is bright with innovations like this and their AI writing tools and enhanced enterprise

[00:29:02] solutions on the horizon too but what are your thoughts on the impact of AI in your business communications how do you see specialized AI shaping the future of translation please share your perspectives by emailing me tech blog

[00:29:17] writer outlook.com X Instagram LinkedIn just at Neil C Hughes nice and easy to find let's keep this conversation going but I'm nearly out of time now so please stay connected for more episodes where I will keep trying to uncover the latest

[00:29:34] in tech innovation and how it generates business value delivers ROI on those tech projects and much more but more than anything keep exploring stay curious and I'll be back here bright and early tomorrow morning so I'll speak to you all then