The biggest tsunami coming is voice: Interview with Richard Delanty, Smartcat’s new partner in Asia

Continuing with our series of interviews with language experts from around the world, this time we reached out to Richard Delanty. A true veteran of the translation industry, Richard has recently started a new business, Into23.

Richard Delanty of Into23

Recently, Into23 signed a partnership agreement with Smartcat to become its official Technology Partner. This will allow Smartcat to boost its sales in India and Southeast Asia, and Into23 to offer tailored tech-powered solutions to its end customers.

The Hong Kong-based company will be mainly focused on Indian regional markets but will expand into other Asian markets as well. We asked Richard about his new venture and the challenges Western language companies face in Asia.

— Can you tell us about your experience as CEO at Into23?

A wonderful journey so far with many highs and some lows!

Every day brings new opportunities and challenges. I have come from leading a team of around 1,000 people to starting from the ground up: but I have enjoyed getting back to the basics — from helping out with our IT setup to helping build out our supply chain to making the coffee but, most of all, getting out to meet customers as a sales person.

We have a big vision for the business but we also recognize that you have to keep your feet on the ground starting out. The key is getting first customers on board, building a repeat client base and then growing from that. It is refreshing to be re-learning a lot of the things I got far removed from and being able to see them with new perspective.

The most rewarding part of it is working with our team. We are still a small but growing group — everyone shares the same vision and dream for the business. I am grateful for their energy, their enthusiasm, their hard work, their passion for making this happen and, above else, for believing in the business.

— You have been living in Asia for more than 20 years already. What made you wish to stay that long? How does Asia compare to being in Europe or the Americas, for instance?

I initially moved to Tokyo in my early twenties with my backpack. I had taken a 3-month contract from an Irish localization company to help set up their Tokyo office. I was then going on a year-long backpacking tour around Asia before heading to Australia. I still have that backpack! I went with few expectations but immediately fell in love with Tokyo.

After 3 years in Tokyo, I moved to Beijing to head up a small office there. Shortly after this, I joined SDL when they acquired the company I was working for. Over the next 17 years, we grew from a small team to 15 offices in 12 countries and 1,000 people. It was a wonderful journey, I worked with some fantastic people, both colleagues and customers. But all journeys must come to an end so I left SDL last year to found a new business.

Asia compared to Ireland: better weather, better food, worse Guinness.

— You’ve been in the industry for a long time now. What is your perspective on the changes that happened during all those years? What were the most significant shifts in Asia?

I moved into localization by accident. I started my career as a developer, or more realistically, I funded my travels by working in development. I started my localization life as an engineer in the company that eventually became Lionbridge, in the Dublin office. Those were fun days but there were few tools, few processes with fewer followed. The biggest shifts are:

The use of technology. I think this is an interesting time for language technology: new cloud based solutions are emerging — like Smartcat — that challenge the traditional license-based on-premise models of before. A lot of the established players are moving to the cloud — but hosting technology is not really cloud-based technology. The challenge for the emerging cloud players will be provable scalability, meeting a deep security audit and robustness at enterprise scale.

Obviously, MT has had a big impact and will continue with moves to neural MT and advances in what that can do in Asian languages.

The other big change has been the nature of work. It’s gone from big distinct projects to continuous localization. You can’t successfully manage this kind of work without a strong workflow technology solution.

From Asian market perspective, I believe the biggest shift is now only beginning. New Asian global businesses are emerging. Look at the smartphone market: apart from Apple, it is almost all Asian companies with the Chinese smartphone companies quickly gaining market share in new markets, such as India. These companies are having immediate and significant localization needs, often into languages such as the Indian languages and other Southeast Asian languages. And also for the usual European languages. But how they are structured internally, what they want from a supplier is very different from the traditional approach localization companies offered. This will challenge and change localization in ways not yet seen.

— What do you see as a major growth opportunity for the industry in the next few years? What about the Asian market in particular?

The growth of localization needs from Asian companies — in Asian and global languages. India will be a growing market. Southeast Asia will be a growing market. More and more Chinese and Indian companies will go global. While global consumer product companies will need to pay attention to the regional markets in India if they want to be successful there — English and Hindi will no longer be enough.

But the biggest tsunami coming is voice: voice will, over time, transform demand for language. In the not-so-distant future, when you buy a new Canon camera or Huawei phone or Samsung watch, you won’t read a user guide, you will ask for information from voice agents: from Samsung Bixby, Amazon Alexa, Google Assistant, etc. Your camera will come with voice support available through these devices but it needs to be across all languages and dialects. This will come from Natural Language Understanding (you asking questions and the voice agent understanding what you need), and Natural Language Generation (the voice agent answering your question). This will require expertise in corpus building across languages as well as some linguistic tagging expertise and some understanding of machine learning. This is the area Into23 is building specialization in.

— What are the biggest challenges the translation industry in Asia faces right now and what are Into23’s approaches towards these challenges?

Today, the biggest challenges I see from the perspective of business buyers are:

None of the big global translation players are really active commercially in Asia. So they often end up working with commercial decision makers in Europe or the US who don’t understand the local business culture. Or using local suppliers that are domestically focused and usually don’t have strong technical expertise.

Translation is not seen as strategic, it is managed deep in the organization in divisions with no cross-company views on translation consistency. Localization companies in Europe and US, particularly the big MLVs, typically compete to win RFPs from central teams. You win the RFP — you win the business for next 2 or 3 years. In Asia, you rarely see this with a few exceptions. So your buyers are fragmented. I think this is why most big MLVs shy away from investing into Asian markets commercially and when they do they fail as they expect the same opportunities here as the Western world can offer.

As translation is not usually a separate buying category it often comes with add-on needs. In my experience, LSPs are build for the traditional translation buyer, their supply chains are for traditional services. When requirements come with unusual needs, they struggle.

Our solution here:

  1. Build relationships. Go do the hard graft of finding individual buyers, get them to refer you to others in the organization.
  2. Build systems. Look at building solutions that can consolidate disparate buyers across an organization e.g. via a portal or an API integration into their systems. Don’t expect them to have a budget for technology licenses though. There is no buyer for that as there is no overall owner.
  3. Understand that translation is rarely what the customer needs: it is a component part of a bigger problem. Have solutions for some of these bigger problems through relevant partnerships.

— On the topic of translation technology, what are your thoughts on machine translation? Have you tested the latest Neural MT engines from Google and Microsoft? And if so, do you have any feedback?

We are developing our MT strategy. Our strategy for using technology is to integrate the right solution in the right place. For example, I think Iconic Translation Machines have a fantastic solution for eDiscovery. Our initial experience is that Microsoft’s Neural MT engine really offers productivity gains for general translation but we are still evaluating across languages. We have yet to evaluate the productivity potential of Google’s MT.

We are also keen to partner with MT providers that need help developing engines for Asian languages — help with evaluations, help with developing and acquiring corpus. This is an area in which we intend to develop a strong core capability.

— Have you heard about Lilt and their approach to MT?

I am a huge fan of Lilt: I think their approach to combining MT and TM is how it was supposed to be done. I love how it removes the need to train MT engines — your MT engine learns as you translate, future output will be more and more customized to your style which improves your productivity. For translators, it should remove a lot of the pain from post-editing.

— Can you tell us more about AI Voice Agents and why Into23 decided to invest in them?

Voice will be one of the big emerging technologies in the coming years, although its capability is still at a very early stage. We believe very few language companies can provide NLU and NLG capabilities today. NLG, in particular, is a fledgling technology. Few companies really have a solution here. Much of what is done today is delivered through templates. As this is not scalable, new and more sophisticated solutions are emerging. We see a big potential in this market and are discussing with a number of potential partners.

— Any other technologies for the future that you are looking into developing or working with?

We have a partnership with CCS Insights. This company provide market information, analysis and intelligence for the connected world, covering the smartphone market, IoT, wearables, mobility and digital workplace. They basically live and breathe all things mobile. We will be partnering with them to grow their customer base in Asian markets.

Our core value as a business is we will build the solution our customers need for international markets. We partner with the best to build custom solutions that work.

— What do you think about our Smartcat ecosystem and why did you decide to become its ambassador and reseller in Asia?

I love it! After leaving SDL, I purposefully put myself through a de-SDLization process. I don’t mean this in any way as a negative comment on SDL which is a fantastic company and where I had a wonderful career that I am very proud of. But I needed to understand what was new in the market, how other companies looked at the challenges in localization.

Smartcat’s approach immediately appealed to me — the technology is really easy to use, intuitive, the onboarding is painless. It has the functionality you need from a translation workflow without being bloated. The ecosystem concept and the payments solution are what really attracted me the most. The open API approach is great — you are providing companies with enough functionality but also a platform where they can build their own solution. This is a massive headache for buyers of translation into multiple languages and Smartcat simplifies it.

Then there is the collaboration capability: allowing translators, editors, customer reviewers all collaborate at the same time. This turns translation from a sequential process into a collaborative one, speeding up turnaround times and helping control quality and consistency all through the process. It empowers project managers with excellent reporting capability, full visibility of progress, simple supplier management and easy payments.

Finally, Smartcat’s security credentials are top notch. For me, this is one of the big questions customers have today — how can I guarantee the security of their data? Smartcat is a secure closed environment with controlled access, encrypted data both at rest and in transit and SOC 1, 2 and 3 compliant data centres.

— Thank you for your answers and insights, Richard! Any final thoughts?

Get in touch with us! Whether you are interested in working or partnering with us, or want to find out more about what we do, we’d love to hear from you.