Artificial intelligence

How Tomedes Brings “Human Touch” to Their Machine Translation Solutions: an Interview with Maya Ronen, Chief of Operations

The US Bureau of Labor and Statistics reported that from 2020 to 2030 that the translation and interpretation sector would grow by 24%. Due to the increasing demand, many professional translation companies have supplemented it with the translation technology available in our market. Recently, some have questioned if the translation industry is over-relying on true automation for translations, causing poor quality?

Today, we will interview Maya Ronen, the Chief Operating Officer of Tomedes, a tech-driven translation services company that combines advanced technology with the expertise of their native translators in their language solutions. As someone who has been part of the translation industry for several years, we wanted to know what it means to have a tech-driven company while keeping the “human touch” in their services.

If you want to learn more about this, then keep on reading!

What is a Technology-Driven Workplace?

In our interview with Ronen, we asked her what it means to have a technology-driven workplace in their translation company. She told us that being a tech-driven workplace means that all aspects of the company incorporate some form of technology not just in the translation process but in other areas of their workflow.

“Sometimes we use existing technology, and other times we develop our own technology systems to meet our needs to make better reports, processes, and services,” Ronen explained.

When we discussed how Tomedes designed technology for more efficiency, she said that they created platforms for their vendors where they can communicate with them. For their clients, they have portals where their clients can have more control over the progress of their translation projects and observe how it’s developing. They created a direct API with the Google platform to improve communication and productivity.

In our discussion, Ronen pointed out that it’s not only professional translation companies integrating tech-driven principles in their production. Efficiently adapting technological trends to increase productivity has led to the demand for the application of true automation, making it more of the norm now than it was a decade ago.

Tomedes Being a Remote-First Company in Its Inception

When 23% of US internet users were still using dial-up, in 2007, Tomedes was founded as a remote-first translation company to provide the best service in the translation industry by combining the expertise of thousands of remote translators with state-of-the-art technology.

Despite the pandemic, the translation industry has seen a significant increase in demand. Experts say that the translation industry will see a 6.7% CAGR growth from 2021 to 2027.

On a smaller scale, Tomedes has greatly benefited from its remote model. It has allowed them to put money that would have been used for a physical office into other aspects, like developing new technology, while keeping their prices at a competitive rate. Remote tools have also made it easier for companies to shift from a conventional work model to a remote one, as communications on such software as Skype and Zoom have greatly improved.

But there are dangers to over-relying on technology in the workplace. Ronen mentioned that technology fails, and it’s normal. She explained that human involvement is still important because machines can’t think outside the box. So it’s up to us to find smart solutions on how to resolve issues.

How Professional Translation Companies Incorporate Technological Advancements in Their Sector

Over the years, Artificial Intelligence and machine learning have become integral not just for professional translation companies but for all companies across different sectors. But the history of AI and machine translation goes way back. Both developments worked side by side.

The first successful automated translations took place on January 7, 1954, at IBM headquarters in Georgetown, New York. The computer translated 60 Russian sentences into English without the need for a translator. At that time, it was called the “Electronic Brain.”

Over time the algorithms used in automated translations that we see some online translators developed into Statistical Machine Translation (SMT). It came from a simple idea: if we had identical sentences in two languages by splitting the sentence structure into words that match the two language words after. Through this method, translation became more efficient.

Recently, there has been much hype on Neural Machine Translation (NMT). This type of MT relies on a neural network system, which means when you encode a phrase, the system decodes it in another language. The decoder only knows its language and can express the data based on its language. Two neural networks work here. One system specifically encodes a set of features while the other decodes it back to text.

Despite how sophisticated MTs have become over the years, they aren’t as accurate in their translation compared to linguists. MTs having an average improvement of 3% to 7% per year. Because of this, many professional translation companies still rely on linguists, as true automation for translations isn’t yet possible.

The challenge of any translation services company is finding the right balance of MT application with the expertise of their linguists in their language solutions. For Tomedes, they claim they have achieved this through their systematic approach to their translation process from Client to MT and linguist and the finished product back to the client for evaluation.

Putting “Human Expertise” into Smart Translations for a Holistic Approach to Customer Service

As a translation company with a customer satisfaction rate of 98%, we had to ask Ronen what their translation workflow looked like. She explained that when a client makes a request, Tomedes will look into the client’s needs. They would try to find out the most important aspect of the translation for their client after going through all the aspects of their request and clarifying certain details. They would then give them the relevant quote.

But it all depends on the kind of request of the client. Ronen explained that they provide Machine Translation Post-Editing (MTPE) solutions if their client needs fast delivery for their translation.

If it’s a regular translation quote, she explained that their team would analyze the request and assess which native translator (vendor) would be the most qualified in handling the project. They will then contact the vendor if they’re available before negotiating their rate for the project. If the client provides them with reference materials, they also send them to the vendor.

Even though they finished the project and sent it back to the client, Ronen told us that they would sometimes make follow-up questions to the vendor depending on the client’s remarks. But even before they send the project back to the client, it has undergone the assessment of the internal operations team through QI checkers and quality assurance before being sent to the customer support team, who will check if it follows the client’s request.

We noticed how different departments were involved in the translation process, so we asked Ronen how important the “human touch” was in their tech-driven workplace and industry?

“If there’s no human involved in the process, it causes problems. If no one checks it or reviews it, that’s where you see problems in translation arise. This is why it’s best to have humans involved before and throughout the process. Machines are limited. They only follow what is directed to them, so you can’t expect a lot from them,” she explained.

From our conversation, the idea of true automation in the translation industry could happen, but relying on it would affect the quality of the translations. It makes sense as each language is bounded by its native speaker’s culture, society, laws, and history. Something that machines still struggle to mimic or understand. It’s also a reason why many courts don’t accept translated legal documents from MTs, still favoring documents translated by linguists.   

What It Takes to Become an ISO-Certified Translation Company for MTPE Application?

In our discussion of certification and accreditation, we found out that Tomedes has recently achieved three ISO certifications. The International Organization for Standardization (ISO) is an international non-government organization that creates and publishes various requirements and standards for businesses and institutions in different industries.

“It’s a very tedious process. Each ISO certification has certain criteria. ISO will evaluate if the company complies. For example, ISO 17100:2015 requires that every language is up to the standards of ISO, and all native translators are also ISO certified themselves. Different ISOs meet different things,” Ronen explained.

One of the ISO certificates that Tomedes got accredited was ISO 18587:2017 for their MTPE solutions. It ensures that any translation company that acquires this certification has an MT application handled by experts who post-edits all translation projects. As mentioned, most professional translation companies use MTs to provide quick translations, especially for projects wherein time is very limited. But the poor management of MTs can lead to low-quality translations.

Ronen discussed that ISO certifications are essential as it ensures that your business follows the criteria of an international standard, making your services accepted across the globe. She said it has helped them re-evaluate their company’s structure and system, improving on specific areas to provide the best service that’s at the same level as the big players in the translation industry.

What is the Future of the Translation Industry with True Automation?

In our interview with Ronen, we brought up some concerns about the future of the translation industry. For example, many people fear that translators will be replaced by technology. It has caused some concerns as some translation issues have been in some sectors like the entertainment sector. This was brought up due the recent issue of Netflix’s Squid Games, wherein many Korean viewers saw several translation errors in the show’s subtitles.

From our discussion, the translation problems of Netflix and other industries are due to the issues of time constraints, quality, and efficiency of production. As said before, the translation industry saw an increasing demand which will continue to grow in the coming years. Finding balance in creating fast and high-quality translation while at the same time meeting up to the deadlines of the client is an issue any translation services company is struggling with.

For Tomedes, they achieved an excellent customer satisfaction rate while giving top-quality translations through their quality assurance in ensuring that every step of the way, their client’s project has been thoroughly evaluated and checked before being returned. Because of this, Ronen believes that people shouldn’t be afraid of true automation and the technological advances in the translation industry.

“The roles of translators have changed. Instead of having just machines or humans alone to do the work, you will have to combine the two. For now, it’s still not there. I think humans will be involved on some level,” she said.

To Top

Pin It on Pinterest

Share This