How Taia helped over 500 businesses go from local to global

Matic Užmah: Okay. I think we are live, so hello everyone. We are live with the other SaaStock event. This time we are streaming on LinkedIn, Facebook and Twitter. And my guest today is Matija Kovač, co-founder at Taia, a UK based startup that is already doing a lot of revenue and recently raised more than a million Euros, I think, in 2020.

So this app is built to using deep learning algorithms, AI automation, and professional human experience enables logic, accuracy and cost and quality of language translation in real time. Matija, welcome and happy to have you on this rant today.

Matija Kovač:  Hi, Matic, thanks for inviting me.

Matic Užmah: I read somewhere, I think it was in the UK tech news, that Taia is more accurate than the Google Translate. So what makes you better in terms of your product?

Matija Kovač:  I don’t think we actually compete with Google Translate. So Taia is essentially evolved from a traditional translation agency where we realized that there’s a lot of space in this industry to automate things and make the entire process much faster and much more efficient. Than this is basically what we’re all about.

So we took how companies traditionally translate their content by sending it via email to translation agencies and getting all the quotes together. And then deciding for this specific service and then waiting for these agencies to do their manual process of sending files up and down and translating them.

We decided to streamline this entire process into a much more simplified solution where a client would just drag and drop their files. They would be analyzed immediately. And once they submit the order, the project immediately goes to the most qualified translator available. So it’s not a Google Translate alternative in most cases, because in most cases, we work with businesses and businesses still require a professional level of service. So they need someone, a human in this case, to verify that the machine translation is in fact, correct.

Matic Užmah: Okay. So you talk about transition from agency to SaaS product. Can you tell us a little bit more about how did you acquire the first customer and how this transition happened? Obviously, you found a need on the market. And then how did you come to the first customer and how did you build the product around it?

Matija Kovač:  Yeah, exactly, just what you said. We found there’s a need on this market. We essentially started out as a language school and the language school is still running, but a lot of our clients were major corporations like banks and so on, who asked us if we can translate their content as well.

So we were basically pushed into the translation area as well. This wasn’t our initial goal, but we realized that this is a very huge market. It’s about three times the size of the entire music industry. So it’s an entire 56 billion USD annual global market and it’s, it’s very fragmented. Not much has changed until very recently.

A lot of companies are still doing things the way they were doing them back in the day 90s. And once we realized that we decided that we can definitely make things different. We can make them faster, we can make them more efficient. So out of necessity, when running a manual or traditional translation agency, we realized that we need a good management system in place where we can automate tasks, we can make things go more smoothly.

We realized that there isn’t such a solution on the market and we started making ideas for one, and then later on developing the demo, presenting this demo to some potential investors and also clients. And we realized that there’s actually a need in the market for what we were aiming for. So we started building a minimum viable product, an MVP, and this one was actually greeted with quite a lot of interests from our existing clients.

So companies who were ordering translation services with us and we realized that yeah, we were on the right track. We should definitely keep pursuing this and make this MVP into a more stable, a more serious solution. Long story short, this is where we are now. So we went from very traditional software as a service solution route, where we first made a demo and then an MVP and then a beta version and then a public beta version. And now I think we’re somewhere around 1.0, something like that, with a lot, a lot of things still to do, so our bucket list is probably endless.

Matic Užmah: So what’s your background? You started the agency, so I guess you were translating a bunch of text before. So what’s your background? Are you in engineering or?

Matija Kovač:  Well, actually we’re two co-founders, Marko and myself. Marko is an economist and a businessman with a wide array of experience. And myself, I studied Chinese in university. And then even before that, I was always very into computers and programming. So I’m the tech side of the company. I don’t deal with languages on a daily basis anymore, even though I enjoy learning new languages.

And Marko is the more of the growth side and administrative side. So the economists in the deal. I’m the guy with the glasses, that’s how I usually say.

Matic Užmah: Okay. So as we mentioned before, you raised 1.5 million. So what was your initial idea? What was your plan? Why did you raise money? Nowadays a lot of founders have on the table – should we bootstrap or are we going into the enterprise or should we raise money, because going into enterprise, you need a lot of salespeople you need more salesforce. So why did you decide that and what was the process?

Matija Kovač:  We realized quite early in the beginning of this company that bootstrapping the entire company is not an option that’s viable for us. Not that we weren’t profitable enough, but the very idea that we would be way to slow this way. It would take us years to generate enough revenue, to be able to fund the development that we are able to fund now.

So we were more keen on growing faster and making sure that we get a larger stake in the market than we would if we were just to bootstrap. The idea here is that we’re in a huge wave of AI taking over a lot, a lot of tasks that humans used to perform or still perform in some way. But AI is taking a lot of this and the companies who adapt and who are able to serve this wave are going to be the ones who will be able to keep going in the next stages of this evolution.

While the companies who are not willing to adapt or can’t adapt because they can’t fund such technological development would probably be run over by this wave of incoming technology. So we decided we don’t want to be the guys lying on the beach getting run down by the tsunami.

We want to be the surfers. And to be the surfers, we need the right tech and to get the right tech, we need a lot of funding to be able to build it because as you might be aware, AI and any modern technology actually costs a bunch of money to build, since you need a bunch of very clever people who can do very interesting things and this costs money. So yeah, we decided to go with funding. So initially we raised a smaller round from a local investor sort of an angel investment. Then we came into a sort of a pre-seed round with around 300 K. And now last year, end of last year we closed the latest round, which I think considers somewhere around this, well, let’s call it a seed investment stage here.

We plan to raise additional funds in the future to support our growth, mostly growth and some part will also go back into the technological development as well.

Matic Užmah: So let’s talk a little bit about the AI. The AI is a really broad term and people usually think AI as a general AI. And I’m assuming you’re talking about the very niche AI. So can you explain a little bit the background. What AI is in your case.

Matija Kovač:  Mainly, we’re talking about machine translation here. So as you mentioned Google Translate previously, that’s a perfect example of a neural machine translation AI.

So basically an engine that’s able to convert texts from one language to another via recognizing specific patterns inside this. There’s a hundred different ways how to do an NMT or neural machine translation engine. But then in the end, it always depends on the quality of data you have. And what we have is a pretty decent quality of data. Since we have nicely labeled data, we know what industry our clients are from, and we know what type of content they translate. And if they agree, we then collect this content and we can build a custom machine translation for a specific company, even. So this is one of the interesting niches that’s been popping up lately.

If you imagine a big corporation, they have a bunch of their existing data or they’re generating a bunch of data on a daily basis, and they need someone to build this into an engine that’s custom-fit just for them. So this is one of the solutions, but then we’re coming into an area of technology these days where it’s not only about translation. It’s also about generating texts, it’s about summarizing texts, all sorts of things that you can imagine that you can do as a human. But in a much faster way. So you can’t write a hundred articles in a minute yourself. Some modern transformer models can easily do something like that. And these articles can be pretty, pretty good. So this is something we’re very interested in as a development team, and we’re really looking forward to getting our hands on some of these tech as well. So yeah. Exciting times.

Matic Užmah: What’s your prediction on what will happen with the agencies like translation agencies, if AI will take over most of the tasks?

Matija Kovač:  I mean, I’m no prophet, but I think it’s going to be very similar than what’s going to happen in the automobile industry, for example. So you will still have people driving cars and some services, but lorries and taxis and this very repetitive driver jobs will probably be extinct because of driverless cars, right. Because AI will take over these jobs and I think a very similar thing will happen in many industries.

Translation will be no different. People who translate things that are very repetitive and boring and keep coming up the same way over and over age and age again. This will be the kind of tasks that will be the easiest to automate. So um, I don’t think AI will be capable of providing a very decent translation of literature, for example, any time soon.

So specific tasks like this will definitely still remain in the human domain. But day-to-day tasks, like let’s say, user manuals and contracts and things that are very repetitive will be automated pretty soon. What we’re aiming at, what we were aiming at initially this was about three years ago, four years ago when we were still outlining the idea of this company was to be a sort of hybrid in between the two. So to provide best of both worlds. One is the human layer on top of the automated layer. And I think we are now in the space where we can provide a very good balance between the two. So on the one hand we can have a pretty decent machine translation that gives you back an entire file translated with your formatting intact and all that.

But on the other hand, you can also decide to go with any of the human layers, multiple human layers on top of that and get a very high-quality professional human translation, ordered from the same ecosystem. So, this is where we are currently. But the more we grow as a company, the better our technology stack, the more we will be into more like software-as-a-service-like solutions and less agency-like solutions.

Since I honestly believe in the next five years, about 95% of all translations that humans do will be covered by AI entirely.

Matic Užmah: I mean, I couldn’t agree more. I mean, we are living in times that AI and technology is definitely really, really fast changing; the way how we live and especially with this COVID situation now in all aspects, the technologies sometimes helping us, sometimes it’s a devil. But I’m really interested in what’s going on in some of the niches in the AI.

Speaking about raising money again, and always going back to what has changed for you when you raised first round and the second round. How did you acquire 1 to 10 customers? What did you do? It was like CEO, co-founders calling the old customers? Or did you use the SaaS technique or Facebook ads or Google ads? How did you acquire first customers?

Matija Kovač:  So, yeah, we’re more or less a B2B company. So we mainly work with businesses and a lot of these businesses that we initially started working with were businesses we acquired previously, either as a translation agency or even before that as a language school.

So I think our first 10 clients were almost definitely companies that we’ve previously worked with and we told them about what we’re building and they were interested in this. So it was a very manual sales direct boots on the ground process. Nothing very sophisticated there, except for that we figured out how we can provide value to these companies. And I think this is all it ended up with. So even if your product is hideous and your marketing is horrible and all of that, if you have a product that provides a consistent quality value to your users, it doesn’t really matter how you market it. I mean, imagine, back in the nineties, Google was a very weird name. And their website was very weird for that time and age. It was all entirely empty, just one search box. But the results that that search box provided was what made Google happen, what went and be the company that they are today. And this is the beautiful example of how marketing and names and logos and all that are not so important as the value that your product provides is. And in our case, yeah, I think we provide enough value to our customers to be able to get their interest very fast.

Initially, it’s always hard to explain what we’re doing because we’re of a weird hybrid mix between Google Translate and traditional agencies. But once they figure out that we streamline their processes, make them much more efficient timewise and moneywise, provide an incredible layer of security compared to traditional agencies, and so on and so forth. When we start talking about all these benefits that they get, we easily get their attention.

I think our problems when it comes to marketing are actually on the other side. We have such a vast array of companies that work with us that it’s hard for us to pinpoint who is our ideal customer. You know, on the one hand we work with global international multinationals, like Red Bull and so on. And it’s, weird because you get into this companies and you get to work with them. Then on the other hand, you have a very strict companies, like let’s say Novartis, its pharmaceutical company, a huge international conglomerate, and they’re very strict about how we work with them. But then you have all these smaller companies that are very good clients, we’re happy working with them. And they range all the way from modern software as a service companies, down to very traditional industry producers, I don’t know, even companies producing very basic machines and stuff like that, but they are a huge consumer of translation services. So for us, this is one of the biggest challenges, in fact.

Matic Užmah: Do you put your customers in different buckets? And then try to calculate customer lifetime value and customer acquisition costs and sales cycle and all this stuff. So you can basically predict, who should you attack first?

Matija Kovač:  Yeah. We try to do all of what you said and, and a bunch more. But I’m always surprised by how weird it is to get any quality data out of this. Because, as I mentioned, we have a very vast array of companies then splitting them up into groups you end up with a vast, let’s say if you split them up into industries, you have, very different companies in the same industry, depending on size and markets they work with and languages that they need translated on. And then once more, if you look into the decision makers and people who actually use our product inside these companies, again, it ranges all the way from CEOs to CTOs and down to secretaries and junior assistants helping, you know, do things in the company.

So it’s hard for us to pinpoint the exact person who is… I mean, it’s a viable problem and I like to discuss it with people. So if there’s anyone out there listening to this, I’m always happy to hear some suggestions, some solutions to this.

Matic Užmah: What’s your pricing? Do you have custom pricing for each company to tailor-made their pricing, or do you have different pricing for startups? Which is probably much easier than enterprise?

Matija Kovač:  Yeah, we try not to. We currently have two major products to simplify things. One is our translation services where we charge per word. And it really depends on the quality of the translation that you want. So machine translation can be extremely cheap. And a translation service with an additional proofreader, an additional reviser or editor, so basically three professional humans on top of the machine translation can be quite costly, specifically depending on language combinations. So this is one part, we call this the LSP part of the company, so the language service provider. This is the more traditional part.

But then we’re now at the stage where we’re introducing some new software as a service solution that will be subscription-based. One of them is coming out in the next couple of weeks, it’s called Catapult. We can talk about this a bit later or yeah, I see you just put the sign up.

So Catapult is actually our first software as a service solution that is entirely automated. It’s a traditional SaaS. There’s no need for human interaction on our end to make this service happen. And what it is – to simplify, it’s, it’s a tool that our translators use to work faster and it’s what we’re now giving out to the rest of the world to use as well.

Since it comes from the fact that technology is now at a level where many more people can be translators than previously possible. Since now you have very good recommendations from machine translation, plus translation memory combination. I know this can get very technical very fast, but basically what it does is it can help more and more ordinary people working in marketing or legal companies and so on translate up to three times faster or even more.

And for this service, the pricing will be the same, no matter what type of a company you are, unless you’re a very big enterprise and you need a bunch of licenses then, yeah, we will probably be able to grant some discounts or something similar to this type of clients.

Matic Užmah: How do you lock your user? Do you have a strategy through locking your user, like Slack or similar tools that the more coworkers or employees you invite, the more you lock the company in? Have you been thinking about something like that?

Matija Kovač:  Yeah. I mean, this was proposed by a couple of people inside the company and also outside the company, so I probably should have considered this more carefully. But I like this idea that more people from a single client can come and start using our product. And I think this is definitely something we will try down the road as one of the experiments to see if we can engage more users from a certain company to use our solutions. But for now we decided to go with the very basic, simple pricing strategy where you pay per seat. And each of these seats is limited with the number of words that a user can translate in a month.

Basically allowing us to close the packages down to different sizes.

Matic Užmah: Cool. Let’s touch sales a bit and especially the future sales. So you raised a decent amount of money and you probably have plans to build a sales team. What are your key metrics and how do you approach sales? B2B SaaS.

Matija Kovač:  One thing we’re hoping to achieve with our software as a service solutions, like Catapult is, is to try and sell as much as possible without human intervention. So we were hoping on building an efficient marketing funnel here so we can get the users to be interested in our product because it provides them with enough value for them to be able to convert. But that said, it’s also an interesting way we already see now because we’re now in a private beta stage and we’re showcasing this to some specific companies just trying to get their feedback, getting them onboarded to see how they interact with the product and so on.

And we already see this as a very good foot in the door type of product. As previously mentioned, we work with some pretty big companies sometimes. And in many chances, we called them up, we start talking with them, and they like what we do, but they’re like, yeah, but we still translate most of the content ourselves. We don’t outsource much. We only outsource every two or three months. We’ll call you back then. And this is where the initial idea for Catapult even came from, because then in this case, we’d be like, well, I guess there’s nothing we can talk about, now let’s close the door and go home. But now with a tool like this, we can put our foot in the door and say: “Hey, but if you already translate by yourself with your internal teams, why wouldn’t you consider using a tool that helps you do this much faster” and like, “yeah, we’re interested in that. Let’s hear it.” And then they get hooked on to what this SaaS product gives them. They also start using the rest of our platform, at least this is the idea we’re having so far in the private beta, and I think it might prove as a very efficient idea in the future. So diversifying our products to make sure we can get into the same clients through multiple different doors or, yeah, something like that.

Matic Užmah: When are you launching the Catapult?

Matija Kovač:  Well in about a month, we’ll be ready to go with a public launch. So somewhere in second half of May we’ll be live. Yeah. Basically in a month from today, something like that. And we’re really hoping to get some cool feedback before that even, but we’ve already onboarded I think over 30 different companies who gave us very decent feedback.

I was actually amazed by how quickly they figured it out and how happy companies were with the value they get from this solution. Cause you don’t know, until you build it, until you put it out there, as long as you just talk about it, people can say many things. But once they have to give their cold hard-earned cash from their pockets into your company’s account to be able to use this solution, then you see that you’re on the right track. And here we went with a very traditional, I mean, I’ll call it traditional by the book more is probably a better way. Just building the product to a certain extent, giving it out there for people to test, asking them how much are they willing to pay, getting them to actually pay for it, to validate the decisions. And yeah, here we are now, ready to launch, ready to rumble.

Matic Užmah: In short, customer development. And you did everything really by the book. Let’s touch on customer success a bit because in the title I think we said how Taia had 500+ companies being successful. And here at SaaS talk, we talk about customer success a lot, because it’s, really the main driver of growth. How do you make your customer successful? And how do you approach your customer success?

Matija Kovač:  One thing our customers keep pointing out when we ask them for feedback and we ask them for let’s say some reviews that we can use in different channels and so on. What consistently pops up is the speed and the quality. So we get their translations much faster without them losing any of the quality that they previously were used to. And this is all down to optimizing every part of the process possible. Sometimes with AI, sometimes just with traditional, if you want to call it that way, web solutions, so just a cloud-based platform that helps you do things much more efficiently. No versioning headaches, no emails up and down, right, no trouble getting your quotes, and all of that. As I mentioned previously, the LSP part of the company is already doing this for some time now for the second year on the market.

And we’ve had some pretty good results. They are usually, they’re very amazed by how are we able to get a quote back to them in under 30 seconds. Whereas with their traditional translation provider, it takes them up to a day or even two in some cases to get a quote back. And then also, you know, the benefits of being able to monitor the progress of your project to be able to have different team members, checking in and assigning projects and so on.

And then also, yeah, of course, because of the AI assistance and all of the solutions that we’re implementing, we’re way faster than any traditional agencies, about two to three times faster than you would be normally. So yeah, sometimes clients are just amazed. They order things on a Saturday afternoon and by Monday it’s waiting for them in the app. They just download it and they’re happy with it, and we’re happy. So, yeah, that works.

Matic Užmah: So no churn problem?

Matija Kovač:  Yeah, I guess that’s not our main problem yet, yeah. I don’t remember any big issues with churn so far. This is one of the benefits when you’re a new guy on the street and you’re doing something that not many people are doing or very few are and your clients don’t have many choices of where they are going to go. Once they start working with you, they’re happy with it. and then they stayed there. I imagine down the road, when we have more competition similar to what we do, this might be a bigger problem that we’ll have to address them. But we’ll see how things move along when we get there.

Matic Užmah: So far, you’re in the blue ocean, so that’s good.

Matija Kovač:  Yeah, we’ll cross that bridge when we get there.

Matic Užmah: So how did your day-to-day routine change from the early days? So you probably had a lot of hats, wearing a lot of hats in the beginning. So you’re doing marketing, you’re doing product, you’re talking to your customers. What’s your day-to-day routine now? You have to lead a team, so what’s your main occupation these days?

Matija Kovač:  Yeah. As you said, as a beginner startup, the founders, co-founders are usually guys wearing a bunch of hats. Yeah, you do everything on your own. You talk to clients, you make marketing campaigns, you build websites, you, program demos in the night. You create all this pitching stuff that you need to attract investors. I mean, even the entire charade of attracting investors is an entire separate story. So even if you’re very good at what you do, let’s say technologically or as a marketing expert, it might be shocking how very different it is to discuss whatever you are doing with investors.

Specifically if it’s technically complex, and business-side complex, as what we’re doing is. But then as you grow as company, I mean, we’re very lucky now to have a brilliant team of a bit over 30 people at this stage. And I’m really excited about how things are coming along, how very, very clever and just brilliant people we’re getting on board and it’s a very difficult process. And one thing I’m happy about is that I’m not the integral part of this process of growing the team anymore. We have our HR expert who joined us a couple months ago in, I think, in the second half of 2020. And it’s just amazing how, we didn’t expect the amount of workload and the quality of the results that employing such a person would change. Previously, it was just myself and Marko, mostly doing the interviews, and seeping through all the endless CVs, and trying to get the right people.

And before you also ask, what changed when you were funded compared to previously. Well for one, getting funded got us a lot of media attention. Which is good from the client perspective, of course, there’s a bit more recognition on the market, but it’s not substantial. But what it got also is the feeling of security that people get when they apply to work with your company and the pool of people applying to work with us has changed drastically compared to previously, since now there’s more people applying. We have a much better choice of who we want to work with. And on the other hand, these people who are applying are usually more qualified, much more skilled than what we would previously get, because you know how it is. If you’re a. 35-year-old specialist, who’s been into the industry for 10 or 15 years, and you’re very experienced senior person in let’s say growth or software development or whatever it is that we’re looking for, you’re probably not going to apply to a startup that doesn’t have a very high chance of success. We, on the other hand, managed to prove time and again that we know how to build things, we know how to grow. We know how to stay on top. And now with this latest financing, we’ve also shown that yeah, we can stay on top for some while still, and we won’t be going down anytime soon. So more quality people are applying to work with us, this is one of the biggest changes, I think.

Matic Užmah: Cool. What’s,next for Taia in 2021? What are the plans for the future?

Matija Kovač:  Do we have another 45 minutes to discuss that?

Matic Užmah: Let’s make an elevator pitch. We have three to five minutes.

Matija Kovač:  We’ll keep on growing, validating our SaaS solution that’s coming out in a month. Probably introducing another product later on in the year, but mostly improving what we’re doing now, making sure that our customers stay happy and keep coming back. And just growing, growing, growing. We are now in the scale-up phase, where we need to grow and show the rest of the world what we’re doing.

So, yeah, it’s an interesting time ahead. I mean, there’s just so much more technology we have to build still. The development team has their hands full. The marketing team has their hands full. And anyone else in the company? Well, there’s no time to slack.

Matic Užmah: A month ago, I think I visited your site. And since then, I’m getting on YouTube and everywhere. So marketing is, doing its job.

Matija Kovač:  This is an interesting aspect, yeah, because we tried to reach out to different people and then see who engages back with our website after they’re exposed to a bunch of remarketing. And remarketing, I think is a good strategy because it’s very cost-effective once you get people on your website, it doesn’t cost much to keep on popping up wherever they go and getting them excited about what you’re doing, getting them to come back and finally book that meeting or sign up for a free account and testing out what you’re doing.

Matic Užmah: Awesome. Matija, any final thoughts? If you are hiring or looking for new experts, please, let us know. Because I think a lot of marketers are listening.

Matija Kovač:  We’re hiring marketers, we’re hiring software developers, we’re hiring all over the place. And translators of course. So yeah, we plan to grow the company even more from the current team, but at this stage we finally made it to where we have most major positions covered and we’re ready to grow and ready to build. And being too small a team is not such a burden anymore, as it was previously. Yeah, keep your chin up. Don’t let the AI overrun you, be the surfer.

Matic Užmah: Matija, thank you for joining us today for taking the time. I wish you all the best to you and your team. And yeah, good luck.