Building a company-wide “Language Infrastructure”: How Kanadevia strengthens its IT foundation to support global expansion

Kanadevia’s global growth strategy: reducing friction from language differences

Kanadevia Corporation (formerly Hitachi Zosen Corporation; hereafter “Kanadevia”) operates environmental and energy businesses that contribute to decarbonization and resource circulation, ranging from waste-to-energy plants to water treatment facilities, hydrogen production systems, and methanation technologies. The company aims to increase its overseas operating income ratio to 50% and is expanding collaboration with overseas teams, including through its Swiss subsidiary.

However, one of the most significant challenges in driving this strategy forward was multilingual communication.

To support safe and consistent translation across the organization and reduce friction in meetings and day-to-day collaboration, Kanadevia implemented DeepL’s product suite. The company adopted DeepL Pro for document translation, DeepL Write for writing support, and introduced DeepL Voice for Meetings and DeepL Voice for Conversations to support both online meetings and face-to-face.

(From left to right)

Kenji Sueoka, Manager, Global Information Structure Management Section, Information And Communication Technology Promotion Headquarters

Lee Jaeim, IT Engineer, Global Information Structure Management Section, Information Infrastructure Management Department, Information And Communication Technology Promotion Headquarters

Kenji Ishibashi, IT Engineer, Global Information Structure Management Section, Information And Communication Technology Promotion Headquarters

The multilingual communication barrier hindering global business Expansion

Could you tell us about Kanadevia’s business policy and the challenges you face in advancing globalization?

Sueoka: Our company has set a policy to move toward an even split in operating income between overseas and domestic businesses, targeting a 50:50 ratio. As we expand our overseas projects, close collaboration and information sharing with our international offices are essential.

However, many internal documents are prepared in Japanese. For locally hired employees at our overseas offices, this sometimes made it difficult to grasp what was currently happening. From an IT department perspective, there was also growing demand to provide maintenance information and internal announcements in both Japanese and English.

In addition, when meetings with overseas teams were scheduled, language-related anxiety could become a psychological barrier and sometimes slowed down communication. These language challenges ultimately impacted operational speed and efficiency.

The challenge was not limited to English. In many situations, communication in languages such as Thai or Indonesian is required. Simply assigning English-speaking personnel is not always sufficient.

Kenji Sueoka, Manager, Global Information Structure Management Section, Information And Communication Technology Promotion Headquarters

How did the consideration of implementing DeepL begin?

Sueoka: The initial trigger was strong support from one of our executives, who asked to introduce the official DeepL product. When we ran an internal survey, we also found employees already using DeepL individually. That led us to conclude that we should provide an officially approved environment where employees could use it with confidence.

Before fully implementing DeepL, how did you handle translation needs?

Ishibashi: We used a combination of file translation tools, text translation tools, and general-purpose AI, depending on the use case. But none of these could be positioned as one standard tool for the entire company. 

Some operated on usage-based pricing models, which made budget control and internal approval processes difficult. In other cases, translation quality was not consistently sufficient for professional business use. As a result, tools and workflows varied by department and individual, creating inconsistencies in quality and workload.

We felt urgency: without a secure, company-wide translation environment, information sharing with overseas offices and the speed of daily operations would be affected. That is why we recognized the need for a unified translation infrastructure that could be used consistently across the organization.

Kenji Ishibashi, IT Engineer, Global Information Structure Management Section, Information And Communication Technology Promotion Headquarters

From contracts and R&D to IT: DeepL adoption expands company-wide

How did you proceed with validation prior to implementation?

Sueoka: In 2024, we conducted a Proof of Concept (PoC) involving approximately 100 users. We distributed DeepL licenses and asked participants to use the tool in their day-to-day operations. In a follow-up survey, we saw a strong response: many users commented that the quality was high and that it was usable for business purposes.

There were very few concerns about how to handle existing tools or how responsibilities should be divided between solutions, and internal alignment progressed smoothly. Security and regulatory compliance were also fully assured, leaving no major concerns during the selection process. As a result, implementation proceeded quickly, and about two months later we moved to a full contract.

How is DeepL currently being used across the organization?

Ishibashi: We implemented DeepL Pro, DeepL Write, DeepL Voice for Meetings, and DeepL Voice for Conversations. Users choose the appropriate tool for the task. By providing add-ins and desktop applications, we created an accessible environment that supports company-wide adoption.

DeepL is used in many scenarios. Contract departments translate agreements with overseas partners. R&D teams translate academic papers and technical literature. International project teams rely on it for communication with overseas offices, while the IT department uses it to translate technical documentation.

Lee: Personally, I find file translation particularly valuable. Even when documents contain a mix of text, charts, and images, the layout remains largely intact. The ability to translate PowerPoint, PDF, and Word files directly is convenient.

With general-purpose AI tools, file translation can come with uncertainty, and it can take additional time. In contrast, DeepL processes files quickly and preserves formatting, so we rely on it for most file translation needs.

Lee Jaeim, IT Engineer, Global Information Structure Management Section, Information And Communication Technology Promotion Headquarters

Removing friction in meetings and conversations with voice translation

How is DeepL Voice being used?

Ishibashi: During overseas business trips and exhibitions, many employees use DeepL Voice for Conversations on company-issued smartphones to support face-to-face communication. Due to company policy, the range of applications that can be installed is limited, so having an officially approved DeepL app available has been helpful.

Lee: DeepL Voice for Meetings is also widely used for online meetings. Since there are few viable alternatives, we see clear value in DeepL Voice as a solution that supports both online and offline communication.

Accelerating global expansion with a “language infrastructure”

How do you perceive the differences compared to general-purpose AI tools?

Lee: Recently, I tried using a general-purpose AI tool during a meeting where four languages—Korean, Japanese, English, and Chinese—were being spoken. The translation speed was slow, and the delay was too significant for a live meeting.

In contrast, DeepL Voice for Meetings displays spoken content in real time as text during online meetings. This helps maintain the natural tempo of conversation and allows discussions to proceed without losing the flow of dialogue. Having the same high-quality translation available in meetings as in document translation is essential to building a company-wide standard language infrastructure.

Finally, could you share your outlook for the future?

Sueoka: Expanding overseas projects and strengthening collaboration with international offices are key strategic priorities. In practice, language differences can slow work down. That is why it is important to establish AI-powered translation tools as part of our internal infrastructure—so language doesn't become a bottleneck.

By leveraging DeepL, employees can focus on applying their core expertise while reducing friction from language barriers. We believe this environment will support our continued global expansion.