| Joe |
My name is Joe Holtman and welcome to Globally Speaking. I’m a Business Development Manager here at RWS Moravia and I have five plus years in the localization industry. In my role, I help match RWS Moravia services to our clients’ business goals to help them go global.
On today’s podcast, Alex Zekakis and I talk about his role as a solution architect, his teaching background and advice for young professionals, as well as trends he has seen in the industry and his optimistic view of the human versus machine debate.
I’ll let Alex introduce himself here. |
| Alex |
Thanks for having me. I’m Alex Zekakis. I’m based out of beautiful, sunny and really warm Greece. I am the Global Solutions Architect Manager at XTM International. I manage a team of about 10 people now. And I’ve been in the localization industry for some eleven years now, mostly on the LSP side; up until a couple of years ago, I was always on the provider side. I’ve been with Sajan, Amplexor and NLG.
As many, I accidentally joined the industry. I come from an IT background. I studied Information Management—totally irrelevant, I just happen to have a brother who’s a linguist, he was a freelancer. One thing brought to another and I’m stuck in the industry ever since and loving it, you know? |
| Joe |
What does a solution architect do and what is the advantage of your role for your clients? |
| Alex |
Generally speaking, you would say kind of like a jack-of-all-trades; as my ex-manager used to say at Sajan: the paratroopers, you know?
So essentially, we’re a group of folks that are technically oriented. We understand the product quite well, we understand technology quite well, but we also have a sense for business. So, we understand requirements, customer processes, you know, we understand how localization flows should work or, you know, challenges that could rise depending on the industry, depending on file types, depending on anything like that. So, our role starts off quite early in the process, during prospecting, even.
We make sure that the product is configured in such a way that it addresses their requirements, that it addresses their needs, and that it actually succeeds in that purpose, in that goal. Moving forward, we are also responsible for training; you know, demonstrating the product, making sure that indeed, you know, we give exactly what is being looked for. |
| Joe |
Alex, why don’t you go ahead and introduce us to your product? |
| Alex |
XTM is a cloud-based translation management system, right? We don’t offer services in any way. We’re agnostic in that regard, and I don’t think we’ll ever move in that direction. At least that’s not a plan for now.
So, we offer a product that is meant for enterprises. So, our goal is to give them an end-to-end solution to manage the localization process entirely within a single system. That includes, you know, workflow configurations, workflow management; it includes project management capabilities, advanced searches, dashboarding and reporting and, of course, an integrated translation editor, right? We have a very tightly integrated, you know, LSP module, as we call it, where LSPs can actually connect directly to their customers in XTM and communicate with one another and still work essentially online in one system.
To summarize that, XTM would be the enterprise end-to-end solution that includes all the components so that the entire workflow is processed within a single system. |
| Joe |
Yeah, it sounds like an exciting technology to be a part of in an exciting time when things are always changed in the industry.
So Alex, tell me what are some of the trends you’re seeing among your clients.What do they need to solve their business challenges? |
| Alex |
Well, I would say, first of all, there is a very steep increase in connectors, in integrations with content management systems, content sources.
There’s a huge trend towards automation in that regard, connecting; and then also automating the pieces that typically require manual clicks like, I don’t know, selecting the languages or defining the resources or crowdsourcing, even. That will be definitely one major trend that I’m seeing at the moment.
An expansion of it is obviously machine translation, which comes into play more and more throughout these ten years, eleven years now that I’ve been in the industry. I’ve seen a massive increase in the use of MT in various ways for various purposes.
But I think the secondary trend here that is very obvious to me is that there’s also a general movement to the cloud, right? People are looking for simplified solutions. People are looking for ways to mitigate potential errors or even leak of data or even, you know, being kind of forced to install products, generally speaking. So, the client-server model, a very traditional model, is something that I’m seeing more and more being pushed out of the equation. But since you asked about trends, it’s obviously something that I’m seeing more and more that, you know, an RFP for example, would be immediately a cloud-based solution. Something I didn’t see in the past. |
| Joe |
So Alex, what I’m interested in is the increase in automations and the use of MT to supplement human work and how moving tasks to machine affects the human contribution. How do you think of the human versus machine debate, and how will humans work with machines? |
| Alex |
That would speak to anything technology, right? So, whenever you introduce technology, you introduce a machine. Even if we go 20 years back when the first CAT tools were developed, essentially, they replaced text editors or even typewriters—in many cases, hard-coded pens and pencils, right? So, machine comes into play all the time, even more so when it is machine translation. So, the debate about machine versus human, in my opinion, is quite void. I don’t see humans being replaced by machines, to be fair. I see machines contributing, especially in areas where previously it was impossible, right?
So, to elaborate a bit on that, the content that is being processed by machines many times is of such a volume that humans would not be able to do that, right? If a human can translate 2,000 words a day and your content that you pass through the MT engine is, I don’t know, 100k words a day, you need 50 translators. That’s something you wouldn’t be able to find, right? Or even pay for. So typically, the massive volumes would be for content that wouldn’t previously be even translated as far as I’ve seen, right?
Then on the second part, there’s obviously the contribution of MT to productivity and efficiency. So, use it as part of a process to reduce the costs and decrease the time to market—of course effectively, right—and still maintain same or high level of quality. In that scenario, I see a strong collaboration between human and machine, right? Translators previously could now become post-editors. Of course, they get paid less, so that’s a debate here, but the volumes that they can process is much higher. |
| Joe |
Right. I think the same exact way. I have an accounting background of my education and the development of the Excel spreadsheet did not take away the accounting position. It just elevated it; they’re able to work with more clients. They were able to change their role and dive into things more for the normal at-home person.
I know you are a university teacher, so I’m guessing this topic comes up with your students as well. Are you talking about this with your students? Do they have any fears? |
| Alex |
Right. I just started this year; it’s something quite new for me. So, I’m pretty happy that it happened; it was something that I’ve been pursuing for a long time. I’m actually teaching at a Master’s program for the University of Strasbourg. It’s a TCLoc; it’s supported by Tekom, so TC World. It’s a Master’s program in Technical Communication and Localization. So, it’s not only localization; it actually also touches upon, you know, the technical communication, content management and stuff like that.
There’s questions always rising around machine translation. So, people that have never translated before and want to go into the profession because of their studies, because exactly of what they’ve heard or what they read online, they have that fear, you know? “So, will my profession actually exist in a few years?” And the answer is always, “Yes, of course it will.” Professions adapt. If you translated before in a certain way, you translate a different way. You still contribute to that. So, humans will always be part of that.
And generally speaking, I try and prepare them realistically for the market in that regard, right? Trying to make them understand what they’re going to be faced with, and obviously MT is one of those. |
| Joe |
And what excites you about teaching? |
| Alex |
I don’t know, I just like the whole concept. First of all, these are online, asynchronous courses, right? So, I prepare the materials in advance and students will follow them at any time. I’ve actually done a lot of courses myself, so I’ve participated in many throughout the years, doing from machine learning all the way to gamification and other courses offered online. And I found it a very intriguing way of improving yourself, if you may, right?
I always found it exciting to be able to, you know, share my knowledge with others or make propositions or even hear what others have to tell me. |
| Joe |
You did mention gamification and I wanted to touch on that because gaming obviously has grown in popularity throughout our lives. It was a cornerstone of probably people our age. And it’s not only for entertainment; it’s for education that gamification comes in. So, it’s becoming a very popular medium to have localized. What are some things that people should know about when localizing gaming in your opinion, or some things to be aware of before you start that? |
| Alex |
The first important thing that I want to mention here is that it matters. I speak for a small country. We’re a country of 10 million, right, Greece, and also a unique language. Our language is only spoken in Greece and in Cyprus, which is a much smaller country than Greece anyways. So, the total population is not that significant, right? We’re incomparable to Arabic- or Chinese-speaking countries or Spanish-speaking countries for that matter, right?
But still, because I follow the news quite a bit, I see that certain companies choose to heavily promote their localized Greek content. Gaming companies, right? Like the likes of Sony, for example: their PlayStation ads are in full Greek and, you know, every game is localized. There’s at least subtitles; maybe voiceover many times.
Well, as a result, there might be other factors, but for example in this case, Sony is the biggest selling company in Greece by far in terms of the console and the actual games. And I have to at least partially attribute it to the huge efforts they’re taking in localization.
I think that gaming localization is one of the more wild beasts to tame simply because they’re so dynamic, right? Games have that nature, as software has generally, especially now that the model of live service is the most common one, right? Like Fortnite or other successful games. Live service is essentially a game that is ever-growing, right? That means that there’s content drops all the time; there’s updates, even weekly, happening to games. These updates have an impact on localization.
There’s also considerations around how do you translate, you know? It’s not simply translating text-by-text. You always have to consider the target audience by age, by demographic. There needs to be a very good internationalization and globalization strategy in place for that matter, right? Knowing in advance that these things will happen, having planned for them, you can also plan for, you know, the content drop, for the continuous localization methodologies that you’re going to follow, because it will take a long way. If these come by surprise, you won’t be able to react to them. And if a game isn’t localized, as with anything, it will start losing value, right? |
| Joe |
Yeah. I would assume the technology becomes even more important when it comes to localizing gaming content. Correct? |
| Alex |
Yeah, for sure, for sure. I mean, first of all, I think that it goes without saying that, you know, the code repository not being connected to a translation management system will already, you know, hinder your ability to deliver dynamic content, right? The fact that you will have a good repository, some different repository of your code, automatically synced up with your translation management system delivering those drops automatically to the system on a continuous basis, you know, and all the users in the chain being notified by the system of the change, of the new content and working on that process without a certain person needing to send an email, needing to click a button or anything like that, I mean, it’s a huge time gain, right?
Most of these AAA, at least, developing companies work across various time zones. So, it might very well be that the content is created in the States, but the localization team sits in Japan. So, you have a time difference of at least eight hours, or something around that, until something can actually start. So, if there’s no automation in place, I think you actually might even lose the game there. |
| Joe |
Were you playing English-speaking games? |
| Alex |
Well, when I was growing up, being at the age of 35, localization wasn’t that strong, right? I grew up playing games in English, and I still do so simply because I’m accustomed to it; that’s the way I grew up. |
| Joe |
Yeah. For a non-English-speaking gamer, the content being localized for them and having their type of humor and the flow in the content is very important. |
| Alex |
It obviously makes a huge impact, right? As I said before, the audience is pretty much expecting it nowadays. First of all, it shows a level of respect to the target audience. The target audience is much more likely, as many researchers have shown, to accept a product if it’s in their own language. Gamers actually feel respected when that happens; they feel attached to the game. They feel that they are part of a broader community because there are communities being formed. |
| Joe |
Right, and in the end, a relationship between client and vendor boils down to mutual respect. So, could you describe one particularly challenging project with a solution you were proud of? |
| Alex |
I was actively involved in a project that was pertaining to the localization of content that was never translated previously; entirely new content. It actually had to do with a cognitive AI for oncology. So, something that had not been done before; like artificial intelligence surrounding medical devices and medical treatment. And we had to come up not only with a process and a plan on what resources to use and maybe even how to integrate into the process, but also deliver all that in a very, very, very short time.
So, we had to come up with solutions, first of all, to the resourcing, as mentioned, which included the application of specialized linguists heavily focused on the medical devices. And then there was also limitations on the tools and the technical requirements of the tool itself. So, we also had to add a certain level of automation not only for parsing, but also for pre-engineering, post-engineering of files in order to ensure that whatever flows out of the translation management system can go back to that specific tool that was, you know, the AI system.
It was kind of the perfect storm because in most projects, you would typically have one big challenge, right? You would have the resourcing challenge or maybe timing challenge or cost challenge, right? Now here, we had all of those. Cost was also something that was, you know, an issue. So, finding a solution that addresses all the points at the same time, effectively breaking the iron triangle that kind of forces you to choose two of the three, time, quality or cost—we had to get all three at the same time. Initially, it was a bit of a disaster, I got to say, so there was a bit of chaos, but there was regrouping efforts required there. And in the end, the project was delivered correctly and on time. It was the first time that such content was actually out there. So hence, that’s why I take pride in that and I think it’s actually one of those cases that I see as the most challenging ones I’ve been in. |
| Joe |
Yeah, it’s times like that in your career that you feel like you really contributed. So, as a seasoned localization professional, what advice would you give the younger generation that’s coming into the industry? |
| Alex |
Yeah. And that’s something I always talk about with the students in the university as well. I guess the piece of advice would highly depend also on the role, right?
So, let me just start with the translator. My piece of advice for them is to be open-minded in terms of processes and tools to use. I think a linguist that is capable of using an array of translation tools has the capability of also reaching a bigger, larger audience, right? Adapting to a new tool should be a process that takes a few hours because fundamentally, all CAT tools share [the] same principles. They work in a very similar way, right? It’s a text editor that is a TM and terminology.
If you can use one tool, you can reach certain customers. If you can use 10 tools, you have actually effectively expanded that, right?
I’ve seen that in action. That’s my brother, by the way; he’s been a translator for some 15 years. He can use 15 different tools at the moment, and he never runs out of work, you know, simple as that.
For, I would say, all the other surrounding professions in localization, I think that the first step here is to be prepared for an industry that is ever-changing and be in a position to ever change. If you do some certain things a certain way today, as a project manager, for example, it doesn’t mean you would need to do them the same way tomorrow. You need to be open. You need to be thinking critically on the processes. Processes are not defined in stone in this industry. It’s a new industry—we still don’t fully know what we’re doing or what the full potential is. And we’re heavily influenced by the evolution of other industries, right? Industries go to the cloud, we follow; we’re not the pioneers. We’re quite a small industry to be a pioneer, but we follow those trends, quite fast as well. So, you need to be able to adapt to them. |
| Joe |
Yeah. I think that’s great advice for future linguists and people in roles like yours and mine.
And to wrap up Alex, I’d like to look into the crystal ball a little bit and in a general way, where do you see the industry going in the future? |
| Alex |
So, I think I will follow the general theme of what I’ve said so far and use that keyword automation, right? It comes up over and over and over. Before I do that, one point that I typically like raising in these discussions is the term AI. It comes up a lot. In recent past, we see it over and over in this industry. I feel, however, that we are maybe a bit abusing the term AI or overusing it to express some certain things.
If you’ve done some certain courses in machine learning, if you’ve experienced AI—I mean, even the basic binary operation of 0 or 1 could be considered AI, but that would be 20, 30 years ago. Now, if you’re talking AI, you need to show something more, right? I don’t like when I see the word AI being used so heavily right now, and then you go and find out more and you see it’s simply, you know, a parameter that drives a certain operation. That’s not necessarily AI; it’s overusing that term in a certain sense, right? So, with that said, the trend is still in that direction, right? It’s reaching AI; it’s reaching automation.
Now, content and volumes…the general trend that we’ve seen over the years is volumes growing, right? More content being localized and more resources being allocated to this, be it a project manager, be it bigger teams, larger companies involved in that process and you name it. So, diminishing return is an effect that no matter what, will occur at some point, right? And it’s a financial term, but inevitably, if you add more people to a certain process, at some point it’s gonna, you know, fall down. So, each person would not be able to contribute as much as before. So, inevitably you need automation to achieve that.
Now, that automation that I see being on the rise is that of automating the project creation, management and control process, right? So, the execution of the project is kind of in focus at the moment, meaning that, you know, what content is sent for translation, selecting that content potentially, you know, understanding which content would need to be localized—that is definitely an AI operation that is feasible at some point.
Then moving into creating and assigning that project, deciding which linguist will translate what, which vendor receives which content. Moving on maybe to other parts of this process, deciding which workflow steps will be followed based on criticality of content, based on something, some other parameters or aspects of the project itself. So, there’s a lot of data that needs to be used, a lot of parameters that need to be considered, but I see that trend growing at the moment.
Of course, MT and all that is still in play, right? But it’s been for a long time now. I think the first time I experimented with MT was ten years ago now, maybe nine years ago. So, it’s been evolving slowly, gradually with some, I would say, strides in the last few years, but the focus on that part of the process is definitely something that I see happening. So, AI is a term existing right now, but I think to be truly using that, we still need to make something bigger, something more in the systems that we use in the processes. |
| Joe |
Alex, thanks for joining us on Globally Speaking today and we hope to connect with you again in the future. |
| Alex |
Yeah. Thanks so much for having me. It’s been a pleasure, it was a great chat and yeah, looking forward to speaking, meeting you, hopefully again sometime soon in the future. |
