Cybozu and DeepL: enabling global communication with Language AI

What you need to know about Cybozu Inc.:

  • Tokyo-based software company Cybozu Inc. provides web-based groupware solutions, such as Kintone and Mailwise

  • Cybozu has recently expanded to new global markets and needed a translation tool that allowed for communication "anytime, anywhere, with anyone"

  • With DeepL’s AI translations, Cybozu has seen a positive impact on communication and collaboration across the company

Founded in 1997, Cybozu Inc. develops and provides team-driven groupware to support collaboration, based on its corporate philosophy: "Building a society brimming with teamwork."

Recently, we sat down with Shuichi Suzuki, Cybozu’s Head of IT, to learn more about the company’s global market expansion and vision for the future—plus, how Cybozu leverages DeepL’s AI translation to improve company communication and collaboration.

Can you tell us about yourself and the company?

I’m the general manager of the Information Systems Department of the Operations Division for Cybozu Inc. The main duties of the Information Systems Department include internal system design, construction, and operation. Specifically, the Information Systems Department consists of three teams: 

  • The team that arranges PCs and smartphones, arranges equipment, and responds to requests from users

  • The team that introduces cloud services, such as DeepL, and operates their account management

  • The team that is mainly in charge of development

My role is to organize all three teams.

Cybozu has two businesses: groupware products and teamwork enhancement solutions. For groupware, we develop, sell, and operate four products: Kintone, Garoon, Cybozu Office, and Mailwise. As a developer of groupware used for sharing information among teams and organizations, we are committed to our purpose: "Building a society brimming with teamwork."

Shuichi Suzuki sitting at a desk with a laptop

Can you tell us about your recent global market expansion?

We’re developing Kintone for overseas markets in various countries and regions, including the USA, Australia, China, Taiwan, Malaysia, and Vietnam.

We’re also hiring people regardless of their nationality or mother-tongue—recruiting overseas members in Japan. This is based on the idea that, if we’re going to create a product that will be used in diverse markets, we need to understand them. For some development teams, whether a candidate is a Japanese national or speaks Japanese doesn’t matter anymore.

What were the areas where you felt the use of AI translation and machine translation was necessary?

We have a global presence, with offices in Vietnam, China, and the USA.

At each location, there are more members from each country than Japanese-speaking members. So the language of communication differs depending on the location.

In Vietnam, communication is in Vietnamese—in the US, it’s English; in China, it’s mainly in Chinese; in Japan, it’s Japanese. However, when viewed as a whole, the proportion of Japanese-speaking members is inevitably high. Therefore, much of the communication on the groupware is conducted in Japanese.

Thus, many non-Japanese team members weren’t able to read documents and found it difficult to participate in discussions. This is why we started to consider translation tools, as we felt that we needed to provide support that would remove any language barriers.

There were many options available, including AI translation tools and machine translation. What made you choose DeepL?

The most important thing was that we knew from ‌publicly available data that DeepL’s ‌security standards would be high. I think it was more than four or five years ago that we first considered introducing a translation tool. However, most tools at the time were using the input data to improve their translation systems.

DeepL, on the other hand, has always stated that it promptly deletes input data and handles it securely. As an Information Systems Department, we’re more concerned about security and ease of operation than usability. We started considering DeepL as the tool that seemed to be the best in terms of security.

Shuichi Suzuki standing in his office

How do your users rate DeepL?

Many users say that the translation results from DeepL are more natural-looking than those from other companies.

Other translation tools are said to provide translations that seem blunt or sentences that sound like example sentences. Another problem was that users were unable to judge whether they could use the translation results without editing, and they had to be careful. 

For example, they needed to translate back or look up each word. DeepL has a high reputation for translation accuracy, and the translation results can be used as they are. As a result, many of our employees now use DeepL. The other thing people say is that the desktop application is very easy to use. It seems that everyone uses the shortcut keys to immediately start a translation.

We have a lot of engineers involved in development, so it's also good that there are shortcuts and that it's easy for developers to use.

What do you usually translate and what are the most common situations where translation is needed?

I use DeepL to translate specifications and email enquiries.
At the moment, we’re providing support for products for overseas markets, so we use it for some of those translations as well.

In our company, all internal communication is basically done on Kintone.
We also use DeepL to translate manuals and FAQs created in Kintone into English.

Additionally, we use the browser extension to translate entire pages.

You already had many people using DeepL across Cybozu. What challenges did you face during the implementation process?

Before company-wide implementation, we had several teams using paid DeepL subscriptions. Additionally, almost everyone knew what DeepL was, so there were not many major challenges regarding implementation.

Our internal rules required us to use the paid subscription DeepL for security reasons, but we left account management and contracts to the users. This time around, the Information Systems Department put together a company-wide rollout with a policy of granting licenses to members who wished to use them.

In terms of management, we were able to introduce the system in a way that reduces the burden on the Information Systems Department while also providing significant benefits to the users. This was a major point of the project. For example, SSO allows users to self-sign up and create their own accounts, and when the account limit is reached, the system automatically deactivates unused accounts. 

Shuichi Suzuki standing in front of Cybozu sign

We sometimes hear that it’s difficult to use tools that do not measure ROI, but did Cybozu consider ROI when implementing DeepL company-wide?

Rather than ROI, DeepL is being introduced this time as a solution to a visible challenge.

In this case, the language barrier is an issue that is increasingly being felt within Cybozu. As the number of members whose first language is not Japanese has increased, there are people who cannot understand the message unless it is written both in English and Japanese when it is posted to the entire company. 

Also, there are cases in everyday communications where discussions cannot be conducted unless English is used. The Information System Department is introducing tools to provide support so that everyone can understand communications.

In terms of ROI, I think it's not impossible to show how much time spent on translation and communication has been reduced. I certainly feel the effects.

However, what I feel is the other effect of the introduction of DeepL is that, at the moment, it has made it easier for overseas team members to reach out to me. When my comments are written in the member's native language, the content is understandable and the nuances can be conveyed. 

It makes it easier for everyone to consult with me, and I often feel that the so-called language barrier has decreased. It’s difficult to know whether this is true or not, but I think that DeepL provides a positive impact on our communication that cannot be quantified.

What are the results of removing language and communication barriers, and what is your vision for the future?

From an IT perspective, we want to reduce information and location disparities as much as possible. When a company is trying to compete globally, it’s often said that it’s important to be able to work anytime, anywhere. However, what I feel recently is that "with anyone" is also becoming more important for our company.

After the Covid-19, many companies began working from home, and it’s no longer unusual to be able to work anywhere, anytime. So what comes next in importance at Cybozu is “with anyone.” The first problem that arises when working with a diverse group of people with different ways of thinking is the language barrier. I think that if that barrier is eliminated, a new way of working will become visible.

By sharing information and communicating globally, we can develop new ideas that we would not be able to come up with if we only communicated in Japanese. Cybozu is entering the phase of "anytime, anywhere, with anyone." I think the first step towards promoting this setup is to enable people with different languages to communicate with one another.

Want to learn more about how DeepL’s AI translations could take your global communication to the next level? Talk to our Sales team to find out more.