The Localization Secret 90% of Game Devs Get Wrong

Only 18% of the world speaks English, which means if your game is English only, you're missing out on four out of five potential players. In an increasingly difficult, competitive, and monetization landscape, what are easy wins to increase reach and revenue? And the answer is localization. Sounds boring maybe, but I don't consider money and potentially even the survival of your game or your studio that boring. The tricky part is how do you do it right? What are some of the tips and tricks so that you can do localization in a way that can save you time and money and in a scalable way? So today I'm speaking to Alexander Morowski, the CEO of Alenos, who has been in the localization game for over 21 years and has worked with hundreds and hundreds of different studios in localization. Learn from an experienced veteran that

has been doing it for a very long time right here on Game Makers starting right now. All right. Hi. Hi, Alex and welcome. I'm really glad to have you here with me today. Today, we're going to be talking about game localization, a topic that a lot of people probably put off for too long. And the question I wanted to start off with, Alex, is basically there was an interview with Mr. Beast and he talked about how massive an of an impact localizing the Mr. beast YouTube channels to different languages had on his growth and you know the even greater success of his business. And so I thought we'd kick things off by talking about this impact the economic impact. I I think in today's day and age when we're talking about efficiency and ROI and things like that, people want to know what what is that economic impact. Could you kind of characterize how you

think about it? So localization to one or several or even more languages, what does that impact look like for game developers?
Yes, absolutely. Uh great question about e econom e economical impact of localization and u I know this case with Mr. Beast because my three of my kids they are big fans of Mr. and they watch the um and they they were watching him in their native language and now they they are watching them in English and u what uh he did is uh really great thing by subtitling and uh voice overing uh his video video clips. Uh what he got is a huge audience. It's uh if you look at uh at the numbers only 18% of people in the world uh they speak English. So every four or five uh they will not be using uh a game or product

or uh watch some video without u if it's only in English and if they do not do not understand it that so uh and and if I speak especially about such countries like Korea, Japan and China uh there is 90% of people they do not accept uh products uh which are not localized. they simply don't like and uh they don't don't understand. It seems uh to to them uncomfortable and uh obviously the by localizing you get the audience you get more people more players all over the world and uh check for example you translate into into uh Spanish you get 550,000 uh speakers of Spanish language it's a it's a huge it's U it's more than uh the population of the United States.
So just one

like 550 million. Is that is that what you mean?
550 million, right? Yes. More than the in the in in the US. Okay.
And the Chinese language gives you one one more than one billion, right?
And Hindi is sixund million uh of new players, new users. So and uh they can buy a product and that's the economical uh the the economical effect of localization. So some uh firms some game developers uh they think of localization as a gross hack. When they uh just localize they get more traffic from uh from the stores. Uh so uh if you want to receive more and more players organically from App Store, from Google Play, from Steam, uh if you're not localized, uh for example, App Store would not give you the opportunity to be presented in some local store. They will check that your game is localized into Japanese
and if it's not localized, you will be

not accepted. So you just cannot go there. And uh just not localizing just limits limits this organic uh opportunity to drive more more users to your game. This is simply as that. uh speaking about smaller games, smaller apps, uh uh it's not so it's not so uh expensive to localize into like 10 languages that are top uh that are uh that cover 80% of uh population in um all over the world. So uh if you localize like Chinese, Spanish, English, Hindi, Portuguese, French, Arabic, Japanese, German and Russian, you cover 80% of population on the earth. So it it worth of um just localizing at least first uh user um

interacting screens like game descriptions uh like screenshots like first levels of games.
Mhm.
So that that's kind of thing. So localization gives you uh scale.
Right. So just to understand your perspective, the what the way that you're thinking about localization is really about broadening reach. And obviously as you broaden reach, you also potentially broaden the audience that can that can monetize in your game. And you're saying 18% of the world is English speakers. And so when you think about and you mentioned three or four different languages that you could add to dramatically expand that reach, but how do you typically make that call in terms of
okay I want to add Hindi or I want to add you know Spanish or some other language. What kind of analysis do your clients or the the clients that

you typically work with? How do you think about that? And then how do you actually decide okay here are the three or four languages or the you know 15 different languages. I mean where where do you draw that line? How do you how do you make that decision?
Yeah. So that's a very good question. Uh um just localizing is uh not enough to be successful in uh in a region uh you are intended to go. So uh it starts with uh your marketing uh possibilities. If uh you want to localize into Japanese for example uh even if you're quite a success successful game you will need uh support uh just customer support uh community support. without that uh localization uh will not uh will not help a lot. I am in in in in

some cases where uh this uh this is important. Uh so uh um if you uh want to localize you uh first first um the first thing uh you must think about how we we will uh drive traffic uh for for your game. And if for example the uh advertising uh the cost of us uh user cost of uh player acquisition is too high then perhap maybe localization is not worth it. It will not have the the return on uh positive return on investment. So sometimes uh you'd better localize not into Japanese which is very high paying country but uh advertising is is uh too expensive. Maybe it's it's worse of localizing into Indonesian lots of

people less uh less paying but the uh the advertising cost nothing. This is just an example. So uh what you do what you can do in marketing uh after that comes uh this uh the the selection of the language. If you have some representation uh in uh in Japan in Korea you uh you have some marketing people there you have uh opportunity to advertise there you have some local partners then the localization goes automatically as a need and u that's the best scenario first you plan your uh promotion your marketing your uh opportunities there and then you localize But uh that's the ideal uh scenario. uh the basic scenario is to just to uh translate into 10 languages uh which are if it's uh if we speak about Europe it's it's the abbrevi

so four languages and then C c CGK Chinese Korean uh and then uh you can add whatever you want like Arabic which covers a lot of people worldwide but requires technical things because the layout is right to left not left to right. So it just uh implies that some development uh should be done. uh so uh to to decide uh to decide uh it's uh it's easy first your uh possibilities to market on the market and then just selecting the most popular languages among internet users maybe uh we can uh look at uh the GDP numbers of uh the regions of the countries just to assess the how how uh

they will pay And Alex, maybe I could pause you for just a second there because I think there's a little there's a little bit to unpack in terms of some of the comments that you made. So I I guess the first thing that you talked about which I I think that it would be good for our audience to understand is that when it comes to localization, it's not just in-game, right? So you're talking about what are all the localization touch points. So there's in-game, but you're also saying uh the app description. So app store also marketing also you talked about community so you might have whether it's your discord or your website um so it's not so when you're talking about localization
point one in terms of some of the things that you discussed is that there are there are multiple localization touch points that you need to consider. Second, going back to the original question which is around how do you make that decision about which languages to uh to localize to I I kind of heard two

things from you. Uh the first thing that I heard was that you're saying that you want to actually try to understand the attractiveness of a region. So even though you're saying like Japan as an example may have extremely high ARPO um you know what is the potential ROI for those countries. Um, one thing that I generally do when I look at attractive geographies is to use a service like a app magic for example to check the data for similar games and and so I'll I'll check I'll do the countrywide split and look at those countries but in terms of analysis are there are you know how do your like one thing you said is just go with the base 10 10 languages. So, do you do you recommend that your clients do the base 10 languages because you just kind of don't know or do you actually suggest that you do something like like I do, right, which is like looking at the data

of similar games and trying to trying to determine the attractiveness of those countries based upon other games that have more penetration in in specific geographies.
Yeah. So um regarding those 10 languages, they are uh they've been uh with us for 20 years. Nothing changed. So 20 years ago, the advice was those 10 languages and every uh game localizer would say that just go with those 10 language and and this uh this gives you maximum maximum impact. Uh but if you are very sophisticated uh and uh uh a game developer with uh who cares about unit economics uh uh about uh uh maybe you have investors that that are looking for this return on marketing investment and the advertising you have different channels you want to assess whether it's it converts or not then indeed you you go with App Magic or

better for games and app stores. it's sensor tower or app any or similar web they have all those packages and so you you dig into the numbers but that's not about game development is like the the task of chief financial officer not an indie game developer
right
so um
okay and sorry and and there's one other point that I wanted to um to talk about based upon um what you were previously uh talking about which is that you also suggested that before you localize into a country, you might also think about the investment required because it may not just be
localizing to the country but you suggest you may need a marketing team in whether it's like India if you're doing Hindi or Japanese if you're doing Japan. Uh so uh what have you seen or like if you do localize to one of these different countries um what what do you

see what do you recommend in terms of that on the ground presence whether it's for a marketing function or whether it's just for localization
well I would definitely recommend uh having some uh people uh who would be responsible for marketing in each region
okay
that's uh that's a must but uh that That's uh if if you speak about uh like larger studios uh they ship their game and uh they need uh to work with those all those influencers uh with uh advert advertising in in different markets. So it's a it's a big marketing uh marketing department or they have some uh some marketing agency who is uh do doing the thing. For for small uh indie developers uh the resources uh the opportunities uh are very limited. they they can only just publish their game uh on Steam for example, select all the languages Steam

supports and uh provide all the uh content uh localized into those uh different languages, different screenshots in different languages, uh game descriptions. And this is pretty much pretty much uh um everything that uh that is available for uh just uh smaller smaller studios. Uh they just uh uh spray and pray. I would say
that sounds like most of the game games industry. Yes, we work with different uh game development studios to localize their games and they're very different.
Uh 90% of them are just uh this kind of we will localize at uh to every language we we just uh we we can't just to cover every everything everywhere everything and everywhere. And uh sometimes it works uh it works because uh people on

the app store they just download more and this uh this pays out. But there are some uh companies that are driven by uh finance by by the investors by the return market investments. So they they test waters first. So they for example they auto translate with Google or with Chad GPT some of the some uh some part of the content.
Mhm. and they publish it. Uh take for example Indonesia. They're testing Indonesia. So they uh translate it and then uh and and published and they see that it it it has positive return on investment and so that they they then then they redo the translation by uh just um with more quality and they add some uh representative to uh make this um community support. uh they might do something more in uh Indonesia to to get

some adver some get some coverage more coverage more uh influencers that would be streaming their game there uh locally and uh so they start with uh uh like stepping on a very thin thin ice. So there's testing worries with uh with the localization just localizing
very few of the assets just starting to see
uh and also there there are some uh
uh those um AB testing platforms like similar uh like um split metrics
so you can't AB test how they download or they uh how they install if it's localized or not localized and it's obviously uh when it's you can uh before you invest In the full game localization, you can invest a little bit to just see how it will work.
I would probably consider myself as not deep enough as far as localization is concerned. So I I definitely would like

to ask you some questions about that because it does seem there is more to localization than I had originally considered. But I would say um maybe the question I can ask you because the way I I I can tell you the way that we're doing it which is probably not good enough which is we're we're kind of delaying localization more to the back end right where it's like okay we're oh we've got all these game development issues and problems we got to focus on the game development and then and then we think about integrating localization kind of towards the end and it becomes like I wouldn't quite say an afterthought, but I I I think it would be good to get your expert opinion as somebody who's been, you know, who who's focused on this and who knows the best practices. What is the best timing? What is the best process to maximize efficiency and reduce time and cost?

Um what what do you recommend versus just like uh you know focusing on the game design? Oh, we're going to launch. We're going to launch. Okay. Uh, oh yeah, we need to do localization. Okay. Um, what do you recommend?
Uh, yes, I definitely uh and basically this is my best advice for every developer of whatever a game, app, a website of of whatever. Uh, so u localization should be started uh day one just
oh okay. If you think that your if you have ambition of localizing your uh game into even two languages maybe three languages or maybe 30 languages. If you have this idea that in the future you will will do this then uh you should start with two languages. one uh might be some of your second language but start with two and uh by starting with two just the day one

will imply that you uh that your game will be uh technically internationalized to be uh be be ready to be localized into not two languages but 20 languages because once you start with two languages you think about how the languages will be switched between between each So you get the language switcher and then uh you implement technically how the language switches in the game. So you're using some SDK and you think you do not hard code your uh localization strings into the code. You extract them to to be them in a separate file because you have two files now
right?
Uh so and two is not is not very it's not um uh it's not too much to support uh in in parallel. So even one of the those languages could be artificial or uh machine translated something
but this gives you from the very beginning this uh uh development hygiene

uh so uh you uh you're developing in the way to support those two two languages at the very beginning and then u you get the right plugins if if you're talking about unity how they or right SDKs how too. So you build the process um and you live with localization from the day one. Uh you think about daytime formatting for example, number formatting. Uh the uh SDK that the language itself you're writing your game in uh will be uh just helping you to to um to intentionalize your code. It's in the guidelines of if you developing cotlin for I mean uh Google play. Well, uh there is internationalization guidelines of how to do localizable applications. You'll be thinking about currencies. You will be thinking about uh implementing uh maybe some payment processors if you think about different

markets. Uh and also you'll be thinking right away about the fonts. Uh well, it's that's tricky. uh um for example what's the common mistake uh when the the game is developed by Japanese or Chinese uh developers uh they have hieroglyphs and um the font size and the font the support uh they um the font size is bigger and then in uh if you just use the font in uh Latin uh Latin based world and so uh the size of phone is is not very big problem. But when when western uh developer tries to uh localize their game into Japanese and they have developed some super pixel font that is very immersive in the game but that font it contains only Latin letters. So when they switch to Japanese, no phone supports so much uh

so many uh symbols uh very custom uh very rare custom font supports everything. So uh um knowing about this this uh just uh implies that uh you switch the phones you see how how it will work and then uh if you ever localize to Arabic then the layout uh the layout is different. So it's it's right to left in in Windows. The this uh start button uh was at the right. I was very surprised to see this not on the left. So um and uh just starting with two languages uh solves you uh the whole uh 80% of localization in the future because adding more languages uh costs just adding more localization files uh to your localiz localizable assets folder and that's it. And uh then uh when you have two languages you you become uh bored and tired of managing

them because adding uh the one line of um in some dialogue just one one line or one dialogue uh would imply that you will uh uh just insert those uh texts into two files. It it it's it starts to bother you and then then you come up with a as an engineer with some solution which we call localization platform
that um that uh allows uh it bas simply is integrated to with your GitHub repository. It takes your source language uh um string resources file and then uh allows uh to duplicate it into whatever number of languages
uh translated automatically automatically and it merges all the um all the changes. So you only work on source file in your GitHub. You just uh

make this uh merge and then you get the pull request from the localization platform with all the translations and as a developer you you just don't have any pain of localizing uh just forever. uh it just works. And then uh if you're super uh s super development operations specialist and you uh set up your continuous delivery uh nightly builds, automat automatic uh uh building uh um so you just integrate localization into your build pipeline and you get your game localized almost like automatically. Uh if you add automated translation then it then then it's just instantly localized into whatever kind number of languages. And if we speak

about adding people they can uh human translation they can work on the localization platform and localize uh just uh the next day your updates. So you get your localized build.
Mhm.
Uh almost in no time. And that this is uh how it must work. Uh ideally you you don't postpone the localization, right? You make it continuous. And uh this is uh uh my best advice is to start just with languages and this resolves everything in the future. ju just reinterpreting kind of some of your comments bas basically your argument which I agree with is that it's better to think about localization from day one from the from the perspective of um localization scalability right so for example if you don't consider localization in terms of maybe the code also in terms of like art assets if you bake in images with with text in it you don't want to do that because then it's going to be harder or even in the UI

right? If I in terms of like how you're presenting the language with the UI. So if you consider that from the beginning then later you don't have to go back and refactor the code or redo your the way that the art is done. Um it would be a much more scalable process and much easier to manage in the future. Um,
okay. And and so like sounds like this is a problem that maybe uh a number of companies could be facing now. And when it comes to the the point you made about fonts as well uh you know and I I I remember in the past like one of the big problems for us was like uh with respect to UI was German, right? because we had this much space and we put the German in it's like
but uh but um maybe could you also speak to then other common problems and issues so you know one could be this issue that you brought up which is like not considering localization for for

localization scalability from the beginning but when when you meet with different studios what are some of the other common problems as far from a localization perspective that you've seen what mistakes can we avoid Yeah. So there are technical and then translation errors. So uh technical uh this is your case with uh German text that is not fitting the the size of the button. So over overflowing uh this usually happens but uh this can be fixed by uh the developer themselves. They just see with their eyes without knowing uh knowing uh German for example that something is wrong there. It does just doesn't fit and that's it. So no language proficiency uh required to do this uh QA job. Uh the the most tricky mistakes uh they come uh in localization from mistransations. So uh when when the uh translation is done without context say for example you have a button which is caption cancel

what what so it can be translated in different languages uh in in many different uh time. It might might be uh might have different meanings. Uh whether you are going to cancel or this is just a verb or just a pronoun just just lots of lots of things without a context. Uh the translator would uh translate it in some way and in the game it would come up with in different context differently. And that's the most uh the most uh the most big problem how to uh solve it. So it's structuring uh the resources for the translation. Uh say for example it's not just an excel with with some strings. Uh but it might be a JSON file which has structure like the structure could be like uh game screen number one dialogue

save dialogue uh but save button and the caption on the button is save and you you know just telling from the structure the content the the context of this string and not only the the translator does know this uh but also the the developer also knows that this save is this save. It's used here and not everywhere in the project. So you you may have one button okay and using in in every project but in some dialogue it's it must be translated differently or used in uh so that's a uh that's um that's a point that the most u tricky problem come with mistransations to solve this uh yeah structured string resources and also uh screenshots uh sometimes the localization platforms they allow um finding automatically screenshots that

are matching the the uh strings that are being translated. So it matches automatically using AI or some machine vision and uh so the the the translator would see the screenshots uh because sometimes they uh they okay they they um they always play the game of course but if it's so so big like universe you you cannot you cannot get anywhere. So uh this uh those mistakes mistransations they are um we hunt them we fix them on the stage of quality assurance. So after the game is is localized when we have the the test build then uh we run through all the screens to see that everything everything is uh is fit. So this is uh this is about the mistakes with translations and translations and cont

um and context. Uh also uh the the terms consistency. So say for example you have some characters in the game and they have names and they reference each other and if they are translated different in different ways uh so uh it it breaks uh everything. So uh or there are some artifacts or location names. So once you have some uh axe and uh another in another translation you have a hammer and this is the the basic the same thing and they are mixed. So the glosseries having the glossery having this uh uh book of every term uh every description of every item in the game is very is a must. it. This will help uh translate this u consistently through all the uh through all the story and uh basically this glossery and style guide are the common

practice and when we start to localize game first thing we think and we require from the client to build the glossery. If there is no glossaryy then we build it uh and the proof. So all the items, all the names, every uh everything is on the glossery.
And Alex, sorry, sorry to interrupt, but you mentioned glossery and style guide. How does the style guide different? I I I understand glossery. What do you mean by a style guide?
Yeah. Uh for example, a style guide is a basic description of uh of the game itself or what happens given the context. Uh giving the context for the translator. In the style guide, for example, they they it may contain the uh what how do you address a player? Is it formal or informal? This is very important because it's you or
another you which is polite. So you uh this is this is one of the uh this is one of the uh things they uh usually

have uh in the style guide and also about the translating humor. Uh whether we translate the humor or we don't translate, we just translate or uh do we recreate it or we just translate it with some reference that that was a humor? [Music] Is it literally uh translated or not? So yeah, there are many things that uh give the uh give more of the of the context uh around the game which is uh which is not in the text uh itself.
Okay. So I want to ask a next question but then I also want to circle back to some of the comments that you made because I think it's because you're you're saying some pretty interesting things but it's easy to to to miss because you're just kind of like talk talking through it. I wanted to circle back on those, but um kind of shifting to practical implementation. Um what I have seen a lot of game studios do, especially small game

studios, is just have you you mentioned Excel spreadsheet, whether it's an Excel spreadsheet or a Google spreadsheet. You've got column A, which with the English, and you've got column B for maybe your Spanish, then your whatever, your your Japanese, your your Chinese, your Hindi, your whatever, right? There's different columns. But some of the things that you talked about when you talked about common problems was one one thing that you mentioned was a a lack of structure, right? And so then you talked about well you could have Excel an Excel sheet, but then maybe it's better to have um a a JSON file instead that tells you where the text is actually going within a a kind of structure. I think the second thing you mentioned was this notion of context, right? So, if you have this Google sheet with columns and you don't know where, like if I were to tap on some some like like a specific text string that I want to translate, wouldn't it be nice to see the

screenshot where it's used so you have the full context of of where it's going to be used? So, that that's a second thing. And the third thing was like um that you mentioned was like also having uh the the context of the the the style or how things are are like like the whether it's informal or formal style or or things like that. So with those things in mind, what is the best practical implementation? uh what is a problem that I see so commonly with smaller studios of the Google sheet columns approach? How do we avoid some of the mistakes or problems that you just talked about with like a different way of doing doing things? What what's an alternative approach so that we avoid some of those problems?
Google Sheets is very convenient. It's very obvious way for keeping the localization there. you you have tabs or columns with languages and you have the

the whole the spre spreadsheet. The problem with the spreadsheet uh uh the main one is when you have more than like 10 thousands of lines in the spreadsheet in one language.
Mhm.
Uh it starts to just just sometimes it just doesn't open. Uh it is just just slow slow down. So 10,000 uh to 10 languages and uh basically you it's it's easy to damage. You delete something and and that's it. So it's uh yes there's a version history but you would like to share that spreadsheet to some translators maybe uh that might be convenient that they will corrupt anything. So for very small for for the very start it might work uh but the main the main problem is uh you don't have this.

So every localization platform that we'll be u working with they they will require the key.
What do you mean by common standard? Speaking about the common standard of uh keeping localization resources somewhere is that you keep them uh with key and value key value key values. So uh the name of the string and the string itself
right right
uh is a terrible
uh yes so that's the common thing and if you have key then you already have some structure because in key you you might use dots and this uh you will you will not use JSON but you will use keys and
uh by key you will be uh telling uh that the contents uh the the context of the string uh the developer would know themselves Mhm.
In spreadsheet with no keys, the developer themselves, they do not know where this line uh happens in the code. So this is uh this is just unmanageable

for a small game is it's okay. And there is a terrible format uh PO files which are used on the to translate websites uh where there are no key and the key the source text and so if you change the source text like typo. So you have tippo in in in the in the source text, you just uh maybe mistaken in the source text. You you just change it, correct it and all your translations disappear because the key changes and uh so you your app will no longer see the translations. uh so uh using spreadsheets uh might be convenient maybe for calculations of the the volume of the translation that might be uh maybe sharing with freelance translators but that's that works with with small uh with very small amount of of text uh to get it to get it real

uh the files must be key value or have some structure structure and uh uh the best thing is to use uh some localization platform that will organize all the files in just one place and you forget about just uh managing uh this managing this file.
What is an example of a localization platform?
So there are plenty of them on the market. Uh some of the uh most some of the best are crowd in crowding crowding and tolgall uh there is also phrase.
Mhm.
Uh there is also smart cat. So localization platform uh uh and Google will just uh give you uh so many of them. Um we mostly use crowding in our work and what we like about the localization platforms um at all uh and what the

developers like uh they support every format file file uh you can even work with your spreadsheets why not but uh the spreadsheet will be the resulting file it will be just a container of your uh of your uh localized uh resources But the work on the localization will happen on the platform and the platform will have glossery, style guide, translation memory which is very important and uh different uh machine translation a uh engines and uh modern LLMs to uh to so the the localization platform will allow uh organize the localization pipeline
properly. So uh just uh this is my second advice. If you start with two languages, the next second advice, the next the next best advice is to use localization platform just uh from the very beginning. It's not it doesn't cost

a fortune uh to use it but it organizes uh and um makes it I would say professional. We use uh g repositories to store our code. We use professional ID uh to to develop and uh for localization we should also use professional tools not just manually tossing the the the strings uh in spreadsheets.
Got it.
Yeah.
Okay. So the next question I have is that obviously localization in terms of like you know translating text isn't something that you do on a like a typical game studio would do on a continuous basis. So certainly working with a partner vendor like yours makes sense. Um, but when it comes to choosing someone to work with, how do you how would you recommend that

game studios evaluate different localization partners? Like, you know, whether it it's with cost, turnaround time, quality, what what are the different metrics that you would recommend that uh studios think about when they look at a company like yours or others like yours?
Well, uh that's a good question. So uh there is a universe of uh those who can provide uh localization to a game development uh company and the first the first thing they are the developers thinking that they will have some uh they will hire some freelancers that would do the job and uh uh and for free uh because uh there are some uh crowd uh if you have a community in the game like Minecraft is 80% % translated by the community. Uh so because of the huge uh crowd of people willing to translate uh uh

so the translations just comes uh uh using uh crowdsourcing and uh if you have a community uh of a game uh there might be some translators that would be willing to be credited uh in your game and to translate it for free. This is the first uh thought that that uh that uh that that indie developers uh think of how to uh translate it. But uh the problem with community translation is that this is work and this if the game is big then nobody would translate it 100% because it's lots of work uh which must be paid for. uh there's not so much enthusiasms to to to translate everything. Uh but some translation can can be done. Uh and um but they need some editor, some uh uh managing glosseries, managing

translation memory, building the whole process. So it's tricky. But there are some uh games that are translated by uh crowdsourced people. Then the next step uh is to gather the freelance uh translators
and pay them and they will uh deliver translations. Uh for that you require uh management. So some someone who will call themselves uh localization manager uh which uh will be on the team of uh the game developer. uh sometimes it's marketing or content manager or maybe some maybe a developer themselves. They start to manage uh all those translators. The problem is that uh uh it it's hard to find uh translators that will not will not disappear that would be uh cost effective and uh provide a good quality. um you should be checking uh without

knowing the the target language. You will be not able to check the quality and you will be waiting for your users coming and saying that your translation into French is awful and you just lost your money. Uh so that's uh crowdsource then uh freelance translators your own team uh you just you you just build and uh those are those options are just uh just to to to talk something about about them but the best options of course is to go to some u agency that would do everything. They they will have the linguists of your languages 10, 20, 30. Uh they will have those linguists with the experience of translating games and they uh with the track record and a good translation agency. Uh I would require that they uh dis uh that they uh make everything transparent so that that I know the names of the translators who they are uh they are might be their

linkad in or might we sometimes do this blind CDs. So we make it as as a agency as us we we make it transparent. So uh the linguist uh becomes u a kind of part of your team that will be with your game for many many years. We've we have cases when um we we've been translating the game for 10 years and that was just the same uh translator through all the years just translating into the language. He knows the game. he he's like working uh in in this game developer but just one one uh hour per sometimes months sometimes uh it's week. So the agency would um have uh this pool of translators ready to work and uh the agency would have this localization platform. If you come without localization platform, probably the agency will just insist that you uh just

uh use some and they will recommend and they will help build the process uh and then uh you will get the somebody who is responsible uh for your localization the localization manager on the agency side. So one one point of contact you you will be um discussing every everything uh about localization. Somebody's responsible. This responsibility uh just uh uh is very important because the game developer is no longer responsible for the localization. So the the localization agent is responsible and this frees resources on the game development uh side. Uh how to choose the localization agency? So as I said uh that must be transparent. So uh they might u uh might um reveal uh who the

translators are. their not their CVs maybe but the blind CVs and the names and uh we often have some shared Slack channels with clients with the linguist and with the project managers to just make things faster to fix some something uh um uh the test translation I would say it's if if you have some people inhouse or some loyal users who would be able to check the quality of the translations then I would uh go for uh with test translations within agency. So we do much of that. Uh the company uh they don't know us. They have no trust and they they come and say well uh we have 20 languages. We we must ensure that the quality is good. And what we do is we give them uh as many things are preferential in translation. they give them to choose from the linguists with

their qual uh with their style and u different styles of translation. So if in the in the game uh studio there is someone uh often it's a community manager of uh the native speaker of that that language. So uh if they can assess the the the [Music] quality uh then they would ask uh an agency uh to provide them with the test translation and the the linguists uh maybe some uh different linguists for their uh for their project. So it's building teams. uh agency already has that team uh and it it must be not blackbox. So uh for good agency is not blackbox and uh technically equipped to help uh technically. So that's

what is an example of being technically equipped when you say technically equipped what can you give me an example of that?
Yeah. Uh so uh uh you come with a spreadsheet and you get an advice uh to uh to do that uh on localization platform. It might it might if your technical uh infrastructure if you build uh the way the game the way it uses only spreadsheets that's okay. Uh but the translation must not go uh in the spreadsheets itself. Of course uh you get the platform uh sometimes it's free of charge uh and all the localization work is done professionally on the platform with all those glosseries translation memories quality assurance automated quality checks uh this uh different pipelines when there are two people working one translates the second one is uh approving everything or the machine translates and the second uh human linguist approves

thing. This this part is made uh professional. It it's impossible to do that uh just sitting in spreadsheets and sharing spreadsheets between 20 people that will be doing something. So um I I would say I would say this is what what I mean in uh by uh technically technically equipped.
Okay. So let me just kind of read back to you what what I understood from what you just said. So um I think the first point that you made is that there are very there It's not just about choosing an agency like yours, but there's different options, right? So, one is free crowdsourced, but then,
you know, not every game can do that. Not every game has a popularity of Minecraft. And then you're, you know, it's it's kind of what what it is. The second is hiring freelancers, but then if you have hire freelancers, there's going to be a lot of management overhead. They may not know about some of these best practices. Quality may vary. um and a lot of the issues that you for example have already discussed on this podcast they may not be aware

of. And then when it comes to an agency like yours, what you recommend from the perspective of trying to evaluate an agency is one the transparency, who is doing the translation, making sure that you you really get high visibility into who the person is and and how like the the processes that you guys are using to communicate. Secondly, you were saying that you felt like um I I'm assum I'm going to translate what you said to quality, right? Having a test translation so you know that you're going to get your your expectation in terms of what what the the level of quality that you're looking for. And third, does that agency have the technical support around the tools and and the and and the things that you would expect so that you can do these do the localization in a good way? Is that a Did I miss anything? Is that a good characterization of what what you were say what you were saying?
Yes, the good uh Yesuh perfect. Uh and I would also add about the technical side

that uh the uh the communication cycle the the continuous delivery is very important. So uh the pain the real pain start was uh when there are just two new lines are introduced into the game and the agency uh uh should have an a capability of translating just two lines just next day and two lines in 20 languages means that 20 people will be involved and uh uh these two lines just uh uh without localization platform without the prop proper management this will break all the all all the all the communication all the all the pipeline of u delivery of translation so uh just I would that's very important this continuous support continuous support and also uh what's uh one one more important thing is uh quality assurance this uh stage so uh there are several

gates there are automated quality uh checks in the translated content. Uh that's very ba basic thing. Uh uh and another one is uh uh second linguist who will check the uh the translated uh materials and there is a quality assurance. Uh it's a it's a localization quality testing when uh the linguists uh will test the game uh in their regions.
Mhm. in their uh lo in their local locals with their language enabled. Yeah. And they will check and make uh all the screens and issues. So the full localization cycle is uh so so the localization agency will talk about localization testing uh 100% if if it's quality localization.
Right. And I I have a few follow-up

questions. So, first question is it it seems like we still are talking about in-game localization which uh which you know is naturally what I think about but some of the other points that just circling back to some of your earlier comments in terms of localization touch points it isn't just in game right so it's to your point creative assets for for potentially for marketing community um ASO app store descriptions and things like that so for your customers. What is the how how much of the of your business is in-game versus some of that other stuff and what do you do generally in terms of those other localization touch points? Uh well in terms of uh game development uh 90 I would say 99% of the localization goes uh in game uh in-game localization in-game content and the best uh the best

clients for localization are the games which live for like 10 or 15 years and they have seasonality some events in game. Uh like imagine you have some uh farm and uh to uh to make it not boring uh after a month they invent some Halloween parties or new year or some anything that that uh just uh uh updates the content u the content of the game and uh we like to translate these kind of things and uh so as I said 99% of our localization If speaking about the games, it goes in in games. But when we speak about apps and some SAS or yeah different kinds of uh other businesses, this might sound complicated here. So uh it's um uh sometimes what we translate is more

marketing uh for marketing. So uh uh there is product and product marketing. uh we translate the product and then uh that's a lot of things about u products that is not uh the product itself uh for example help system reference system for software uh a huge amount of text but that's not part of the product I would say uh some blog posts some uh advertising materials some uh case studies uh these kind of but it's it's rare case for for games for for the games um I would say yeah it's what is not the game is this user journey yeah they have very small website the game developers they have a very like just landing page very easy to translate uh and then uh it's mostly app store descriptions

uh and and that's it and that's it and then have they they have their influencers streaming and that's not localizable. Maybe Mr. Beast is localizing his video but uh any streamers they do not maybe with AI or they will do u but it's very limited content that that is uh in in the union in the user journey which is not inside the game.
Okay. Um so I want to ask one more follow-up question to the point on different options or alternatives for localization solutions but I think it'll also lead us into a much bigger discussion which I wanted to talk about which is AI right so like when you talk about those different options crowdsourcing for free I I don't I know a lot of game gaming studios and CEOs and things like that not many are looking at that freelancers maybe a few studios are trying that agencies I know

a lot of studios use agencies for localization, but I think an increasingly popular option for small guys, I'm not talking about the big guys, is AI, right? where maybe you have, you know, obviously to the points that you made, it may not be ideal, but let's say you have a spreadsheet with the, you know, you've got the key English and then other languages and then you just use an AI service like chat GPT or claude or whatever to fill in the translation against the the the English for for example. Um can you talk about that option uh and the use of AI in localization? How do you think about that
and what is the benefit of using an agency? Well, so what one do you use AI?
And then how would you talk to a studio who's like you know what in theory I can just do this stuff with AI? What would you say to them?

Uh well we definitely uh that's a very good question. We definitely use AI and AI um it's it's it's used not only by small uh indie developers but also uh the big guys with huge amounts of content they are they are interested in first place uh because they have resources to uh to choose between the LLMs to choose to train uh to post train their uh the LLMs uh so that uh it it would translate better for them. So uh the uh the most interested uh game developers are the big ones because they have large budgets and they want to cut uh cut the budgets and that's uh
I've never heard of that anymore. I'm just kidding, Alex. uh yeah so uh they were the first who would come to us and uh ask us uh whether we

could translate everything with chat zp and maybe it works but it it didn't in this way uh so uh how it really works now uh you just uh uh you can translate with chdp in localization platform you can turn on the the translation automatic uh automated translation and you will get your translation uh with Charpd or Claude or Gemini or Quen or Deepc whatever uh every PL every uh LLM is supported in localization platform. So you can do this uh and it will pretty much translate well uh for some languages. uh but there is there are some tricks uh not every LLM translates uh good uh in all uh to all the languages. If we uh want to translate into French then Mistral is doing better. If you want to translate to uh

Chinese then deepse obviously does uh that better and not in uh not uh if you speak about European languages maybe CHPD is okay but except French uh to Polish language DL is doing best uh it's not LLM but just the translation uh engine so uh to select the proper uh engine this is what we do uh for our clients. Uh this part of our service when we um when we knowing about the content uh type of the client having the content
uh and knowing about the language pairs we can choose the best AI to do the translations but it doesn't uh translate 100% quality of course it must be uh properly prompted uh promp uh and uh the

glossery is very important And uh the style guide is very important and the manager is very important and uh the the reviewer is very important. The reviewer would check all the translations because you you can go just with uh with uh basic uh machine translation but nobody is responsible for the outcome. Uh as an agency we we are responsible for the quality. So we must check and so we negotiate with the client that we will be translating with this AI for your content type for your language pair. This AI is doing best this those are the tests the metrics we calculate lots of uh technical metrics on the uh to assess the quality of um of the translation. So uh we negotiated with the client that this uh we will be using AI, we will be using a human reviewer and the price would be like 30% less.

It's not 100% less. It's not 90% less. It's like just 30% less because
uh the review would be reading
uh reading assuring the quality and reading and writing that's not very uh big difference uh because back in times where when where there were no LMS there were uh different other machine translation a um engines and also the the translation memory even if this the the uh the translation uh was made in in the past and of the same chunk of text. It must be reviewed must uh must have this seal of quality and this is what we saw responsibility and quality. We guarantee that our localization is good. So and uh for this guarantee uh uh actually the client pays uh this u for for this uh uh for this insurance I would say. Uh so

we are using AI but this is strictly uh when negotiating with the client when they ask or or we can uh just uh roll out several options when they ask us for the price. We usually come with three options just uh human linguist professional linguist who will be translating plus maybe the other editor who will be reviewing this depends on the budget. Uh then uh uh the other option is uh MTP machine post uh MD uh machine translators machine MTP machine translated post editing. Okay. So when uh we agree on that to cut the co the cost uh the cost uh we can use uh for several languages uh uh this or that machine translation plus the reviewer right and that's u and that's those are the two options and the third one is we can go blind machine translation if if

the client has this uh content type for that for example it's a user uh user interaction, some uh dialogues in some chat or some user generated content. Nobody just uh translates, nobody reviews the Airbnb uh comments uh with human because they are just go machine translated and and that's it and that's okay. Uh so uh those are three options. Um and the client chooses and uh they always know that we are using machine translation if if we agree. Uh we never do this just to instead of human human language.
Uh this is uh about the proper way of uh u building building this uh around AI. Uh uh and I I is um yeah there are some

game developers they rely 100% on chpd translations.
Mhm.
And uh they are programming uh the translations through API of open AI and and that's uh that's okay. They sometimes invent some frameworks to check the quality. Um maybe that's okay. uh what what we see we were afraid of two years ago we were too much afraid of this as a business of this scenario but uh recently we see that uh um it's hard to automate to it's hard to automate these translation because sometimes you get unpredictable results uh I mean uh without a glossery your translations will be incon consistent. So you will translating one word in

different ways. Uh this depends on how creativ how how creative how the temperature parameter is set up in in the LM. How creative the LM will be. They will be hallucinating and uh in in inventing something that is not on the original text. This is very common issue with uh translating with CH DPD. And then fixing this uh is uh uh costs money and uh without a glossery the translations are inconsistent but uh when you have glossery it's very big then it doesn't fit to the context window of of of LLM and uh it can be can't can't be used. you you should you must implement the this uh retrieval of part of the glossery insert into the prompt and then translate and then sometimes if you want to translate something about a war bombs or some uh chemicals I don't know there are some games uh with terminology that

CHDP is prohibited to to to talk about and it will say that I would not translate this because it's about some something Okay. So, just just to like just to recap, it sounds like what you're saying is that
sure there are some game studios that use AI, but the advantage of working with an agency like yours is like you guys actually do use AI. You're leveraging it. You know the best models to use. You know how to optimize the use of AI and translation and that cost savings you're passing on to your customers in terms of like a 30% haircut or along those lines, right? and and but then the thing the thing that um the the the area where you guys still add value is that obviously AI is not perfect. There could be hallucinations and if you want to control the quality as well as to have your agency input in terms of best practices and the fact that you guys are actively using AI so you can optimize it. That's the advantage of working with an agency. Is is that a fair characterization Alex?

Yes. Uh you get you get the quality assurance uh uh the quality assurance um of uh the localization and also the testing.
Mhm.
Uh also the testing um of the localized build of uh the game and also this setup uh choosing right uh model um doing all the technical stuff. uh you don't don't need to write the code to use ls for translation because if you want to translate using charg you cannot uh in just insert your JSON file into the charg window and say I want the translation it will work uh for some time uh but if the file is big then you it gets corrupted and if you have 20 languages you have 20 files inserting into chat GPT window and then so it's uh

it's manual work that no nobody likes uh instead of this you will be obviously using API and this requires development and so you're you are just reinventing the bicycle and uh the localization platform in the um localization provider they they will do this and basically it's for free
okay Alex I I thought one other question I could ask you is like if you were to just for our audience so they kind of understand what it's like to work with an agency. Could you walk through maybe a case study or a real world example like when you're working with a client? If you could take us end to end from the very start, what does that engagement look like? How do you interact with them from the beginning all the way to the end? What does that process look like? Could could you give us an example of that?
Yeah, absolutely. that that's again a universe. Every client is uh is

different is different. But uh let's uh let's talk about uh one of the clients that are uh that that been with us for 15 years already more than 10 years. It's long long story but that's my favorite. It's uh uh they started very small. They had a game uh it's a farm simulator. You just uh plant some seeds, you get some fruits, you gather something into buckets, then you make juice, then you you sell juice, you make money, you buy more stuff, more more equipment, you your economy grows. And then there are some characters, some u so this kind of this kind of game, it's evergreen. Uh it has um daily uh con uh constantly uh returning users uh daily users that

are just uh they just love the game and they've been playing for for for 15 years. So uh uh they were just small uh small studio, small game studio and they they have just two languages and uh they started with the US market and that was a hit and uh that uh they had mobile mobile game. No web uh it was mobile mobile game or web. No, that was a web game. That time it was just web uh just web game and uh we uh we got acquainted with them uh on some game development conference. This is the basic way of uh just meeting uh your clients. We started with uh with spreadsheets. It it was far far it's very long story started with

spreadsheets and uh what the from the very very beginning they had a they had uh some translation that were not made by native speakers. That was important. And uh they had community managers uh and huge uh number of users that were playing uh playing the game. and they they started just uh adding four languages uh which was French uh Italian, German and Spanish because they wanted just to to to make uh that more global. Then uh there were the the times where uh the localization platform platforms only just started to appear. That's uh and uh our main value proposition was organizing this uh automated pipeline of uh uh of the delivery uh and and and they uh with them we built this mixed u

I would say I would say the mixed team when uh they are community managers and content managers and our localization managers they are sitting together in one slack channel and communicating like by email by uh u just in chat. uh and uh they have this uh um localization platform projects where they uh they look at the at the at the same thing at the localization resources there and uh from that localization platform the developers they export uh nowadays they export stream resources in different formats. uh they can export as um for iOS uh apps for uh Google uh for Android and for web. So uh this is like central repository u of their

of their um local localizable uh resources. Uh every months uh we have a plan release plan we know the release cycle of them. they plan the next release and they come with uh some files uh they prefer this is the case that when they prefer to create the contents the content in files that's basically Excel spreadsheets because they are writers they just it's it's the long uh it's very very uh long u pro so this is their process from the uh from the ancient times. Uh they they create their content in spreadsheets and then though those spreadsheets go uh to the localization platform. They are localized, checked and uh then returned uh to the client and uh uh two years ago they started to adopt

AIdriven translations. So uh they developed uh with uh wi with with us and uh we developed with them. So they needed more and more languages and we added more and more more uh linguist some exotics exotic linguist and and then uh when they came and asked uh to implement the AI uh workflow, we did that. We built uh um the this uh pipeline of using AI just uh to to help that client and uh it happens so that uh this uh this is just becomes new standard in localization industry to uh build such kind of pipelines in u involving AI uh for the translation. So that was the very right time and we followed uh our client uh everywhere uh when they uh wanted to get some uh

exotic language to uh to do some additional testing to um to help their community managers uh to um mix the teams of two translators. Some of the translations they have inhouse, some of them uh are ours and uh they check each other. So uh we so that's a very long time it's a lifetime uh how we worked with client and um our teams uh they are so our linguists they are like the extension of their team of their development team. So they are players, they know the game, uh they are keepers of the lo of the localization of of that game. So they uh so that that's uh that might be my favorite case. Uh but the common journey

is might be absolutely different. They come with uh the usually they come with just uh asking uh about the price. How much would cost to localize a game of this number of words to these 10 languages? This is the very common thing. We get leads every day of this kind. Uh some studio comes and ask just this 10 languages 2,000 words or 2 million words. sometimes that they're that big and uh they want prices and when they ask uh sometimes uh it tells us that they are very uh localization mature what we say so they uh they know that uh they will get quality they know the all the platforms all the setups they know everything but they need the price uh the only thing they need the price

and uh as I said there are uh always three options and We start with uh the uh the interview uh of the client what they really need, what content types they have. Uh in every game they will have uh different content types and how do how they prioritize uh those content, what goes first, what goes with better quality, what goes with less quality. Um so uh we we build this knowledge uh of and and build a road map of uh the localization the future road map. We understand how uh what the time frame would be and we have this gun chart of just stages uh so so that the client will see how the project will uh be implemented. Regarding the price very basic uh thing it's all uh uh it's all volume based

when we talk about translating translating texts and it's hourly based when we talk about uh localization testing. So um the bigger the game in content uh the the bigger the budget would be. So the prices they vary a lot. Uh one could say it starts from for some languages from six uh uh euro or dollar cents per one word and up to 18 uh for example. But uh it's a it's it never nobody tells the price without seeing the content itself without seeing the game and without knowing the linguists who will be involved. Because when we uh have an estimate of a game uh in terms of price and time frame, we already know the linguist who will be involved. And we see uh if it's the huge game uh of just

1 million words and it will require five linguists and months of work. uh then we could negotiate less prices with linguists and less prices uh to get less price for uh better price for the client. So it it it depends uh so basically it uh always starts with the interview. We understand the client needs and uh uh then we get we we come up with rough estimate uh um pricewise and also time wise and then uh sometimes the clients would require test translations to select the linguist that would be uh best. So we give them three to five uh different linguists with different styles and uh which uh which one is best. They they they're all high quality but it's preferential sometimes. So uh and by by this way we form the

team. We form the team and we just uh uh start the the process. uh we get the localization kit with uh whichever um it might be anything files in Excel whatever or JSON files or already the project on the localization platform uh whatever it might be anything we work with uh with anything
all right great that's uh um you know all the questions I have actually with respect to your business and I thought maybe we could end with just a little bit more of your personal story we're all humans humans on this earth trying to do interesting things and trying to you know uh create value in the world. So I could you talk about how did you how did you get into this business? How did you found your company? What could you talk about that a little bit?
So uh my background is I am a physician and uh software developer. Uh I'm I'm developing to this day some some of the

some of the things at Alenost. So my passion is product development and uh many years ago more than 20 years ago uh I've built a product I just uh I and I started to sell it uh all over the world using e-commerce uh provider and obviously I translated it to some languages I knew and uh uh that was very um uh obvious for me that the software must be in English to sell. Well, uh but I had several uh friends uh we we were uh in the community of software developers that were doing the same thing as me but their uh software were not localized and they asked me to to help them translate uh that software into English and I helped and then I had a girlfriend who translated software into uh German language. So, and I just recommended her

to my friends and they started to order from her and uh I just uh passed the money from them to her and uh every every everything was uh was great until uh the client client one of CL with the clients would come and said that the quality of the translation was not was not good and so I I I must return the money. Uh but uh I just uh that was my girlfriend. Uh she was a speaker of German but maybe not professional translator but who who cared that time. I was not selling translations. I was just uh helping my uh friends to get translations and uh but because of they were sending money to me and I was sending money to her uh they were considering it as as a service and I I that that time I understood that uh if you take money for something you bear responsibility for the quality and

that's goes without saying. So uh and that time uh it just that moment it became serious. So I uh just I understood that if I take money uh and uh deliver something it must be quality. So I I I found another uh translator which which uh would be of more quality. That was a like a little business game for me. Uh not not so serious uh because I I had my day job as a software developer. I had my software to sell uh through the internet and this was the side hustle uh this languages but uh I wrote an article into local newspaper about this experience of translating softwares and suddenly translators starting to uh to ask me to help them to get this uh translation work and I said well okay I give you translation work but I charge you 50% of what client pays

Uh so this is uh this was the margin and uh and this is how it started. I just made an announcement in the community that uh I have four translators into four languages and uh they will help uh you translate your softwares. I will organize every technical things because there were no localization platforms there were XML files files every kind of files and is it easy to broke uh to break it's etc. So I organized this this technical side and communic in communication with the clients and my translators they uh did uh their their their job. And this is how I started just from uh from translating the software myself then adding some uh friends and uh uh it got traction and it it got some traction more languages more orders coming and then uh uh I started to do some marketing uh

some sales activities and uh I got some managers to manage this uh and now uh Elenost was uh last year was more than 100 uh 100 full-time employees which are localization managers.
And uh so the core team uh but if we speak about the translators some of them we call in-house uh they are living in their countries all over the world. It's uh one uh 1,500 translators all over the world. So this big uh this big um yeah
this is this are the company that that grew just from the
right
and and at what point did you did you decide to go all or I mean I assume at some point when you have this much scale you you decided okay maybe I shouldn't be a software developer anymore maybe you should do this business what was your thought process and when did you when did you decide to kind of flip flip

to to this That was there was was a point. It was not um it was not money that uh that made me decide. I I had a good very good salary at my day work as a software developer and also I had this uh revenue from uh uh selling my own software. Uh but I understood that I must build a company because uh as a uh having a day job you are alone. Uh you have this uh glass sailing you cannot surpass some
amount of you you can you can earn if you speak about money and again selling my own software while alone. uh this is a s uh solo developer now it's easy it's AI vibe coding you can do anything just alone that's good but not that time that

time develop development products was uh very hard hard work and it it didn't give me uh much money to hire someone and uh I didn't trust anyone uh because uh those of developers it's hard they were hard to find and you give them the source code why do Don't they just go and sell my software on their own? That was my uh that were my thoughts my thoughts. But uh at that time I somehow understood that I must be a company. I must not be alone. And building company that's a completely different thing than software development uh on on solo. So I uh just knew that I must build a company. I don't know this is my child dream to build a company with some name go with it. So uh when I understood that I'm negotiating with people with translators with u um the developers who were uh

producing software that were much more successful than mine. Uh I I was helping them to localize. uh I uh found my I found myself uh surrounded by people who are doing something and I I um I understood that this is the company is just uh being born right right that time and among my translators uh was one of uh the guys who became a manager and we became two of us
uh and that that's the company and some translator and then uh we just invited someone more more more more and it grew organically without any side investment etc etc and um when I decided to go uh with my own company I I um I had a very good salary but I just I refused to to to to get the the salary I just quit and uh I had no money

sitting in my parents' uh house and I I just was uh taking some money from my mother to pay salaries that was uh the first I mean several maybe even years uh but when while developing alchemos uh I was also I've been also involved in different startups uh I I tried to build something something different uh something different from uh Elconos something else and uh this DNA of doing something else in the company it it's just uh it's just with us uh to to this to this day we are inventing something uh developing some products inside of of the company not only localization. So and the the reason our development is uh how we we just use GPT and Gemini and different LLMs to assess the quality of

translation. We built a SAS service like it's available for free for now but uh it can assess the quality of translations using modern AIM. This is because we uh have the this development culture inside the company. And also we have this Nitra uh uh translation uh service where you just top up your balance. You uh just uh send three lines of text and they are translated into 20 languages by native speaker translators. Just manual uh human translation. this this kind of system we developed uh that on our own and this uh development culture helps us understand how the developers think. So when whenever client comes to us and uh we understand how they think how the developer would organize their uh their files for localization. This this understanding helps uh differentiate. We're not a translation agency which would translate

everything like marketing or documents. We're a localization agency which knows how the developer thinks, how the developer works, what they need. This is the main difference. This is why alenost exists and why it just uh it uh it's been
developed to this.
So so maybe I could ask you one last question along these lines. So it sounds like if if I'm interpreting what you're uh saying correctly that you're developing some type of
uh an AI eval product, right? So evaluating the the outputs of the translations done by AI and that may be productized or maybe not. But when you think about the future of Alenost and certainly you've you've had like this kind of traditional business. We've talked a little bit about AI. AI is certainly impacting a lot of different businesses in different ways. So when you think about Alkenas five years from now or even further in the future, where

do you imagine the company going? Is would do you think you'll have more of a product focus rather than a service focus? Will it be a blend of those? Are there other areas that you'll be exploring? It it'd be great to understand how you think about the future of your company. Uh the future of Valkconost is uh absolutely around localization around languages because uh this is the main expertise. Sure.
Uh and uh uh but it it's not only this service uh it's uh also about the uh uh technical uh it's it's about maybe products around localization and for localization like technical stack for localization. uh we started with ElkenosMT uh project this lab uh where we experiment with AI and um delivering uh different tools for localization they are uh helping us to be on the cutting

edge of the technology. So every new version of uh AI coming we just implement that in our technology stack. Uh so uh I see that uh Elkenost as a uh as a company is a localization services uh vendor. I I would say like uh um and as I like to say we are like a tree. We are organically growing. If there is sun, if there is water, we are just growing organically. And this happens to all the localization companies in the world. We are at a very in a very competitive market. We have 100 companies that we consider and not only me but also different researchers and indices uh 100 companies that are considered our competitors, direct competitors and that's that's a very uh that's a very competitive global business and uh what what we will do is we'll

draw organically and uh disrupting uh from time to time uh different tools platforms uh some features of the platforms. Um I think uh so it's a it's the most uh organic growth. No super big ambition for I've been uh here for 20 years. I've got lots of ambitions and the what worked is uh Elen was doing localization. But as an entrepreneur myself, uh if it's if it could be separated from the company,
uh maybe there could be some side projects, but that what that's that's not um that's maybe different businesses, different uh companies. I don't know. Uh I have plenty of ideas of something else, but uh I would say that speaking about Elenost, it's about localization uh forever.
Got it. All right. Yeah, specialization,

depth. That's all sounds great. Well, Alex, that's all the questions I have. I do want to thank you so much for going in depth in terms of your business and providing more context and information about localization. I do think we probably don't think about it enough and so I definitely appreciate all of the the additional information that you provided today and uh so thank you. Uh if people wanted to get in touch with you, so this is the very last last point. if they wanted to get in touch with you, how should they do that? Should they go to the website? Do you have a social media handle? What what do you recommend?
Well, it's basically link it in uh this place. So, uh if if Yeah, link it in and also amalconos.com. So, am Alexander Morowski amos.com. So, that's uh that's it.
Okay, great. All right. Well, for those in the audience that have made it this

far, thank you so much for your attention and hopefully we will catch you next time. Thanks everybody and thank you very much, Alex.
Thank you. Thank you. Thank Thank you, Joseph. Thank you. Thank you for having me. [Music] [Music]