One persistent challenge for multinationals is addressing the language barrier among their global staff. From instant messaging platforms to official documents, incorrect or inaccurate translations can spell disaster for any enterprise.
This is why global communications platform DeepL has invested heavily in its language translation technology. According to David Parry-Jones, the company’s Chief Revenue Officer, enterprise translation needs are far more complex than those of individual users.
“Effective communication is vital for businesses, whether within the company or with external customers and partners. Using the right terminology in the right setting is crucial. This is especially challenging for businesses operating across languages, given the wide range of terms used across industries and the nuances of language and cultural differences,” he said.
Breaking down walls
DeepL’s translation platform is used by organisations in industries such as tech, retail, and manufacturing, where accurate translations are critical. Its specialised large language model (LLM) technology, developed using proprietary data and refined by language experts, is designed to capture context and nuance across different professional settings.
“This enables us to provide context-aware translations that are accurate and human-like across a range of industries. Additionally, DeepL offers personalisation features, including a glossary tool that allows professionals and companies to customise translations for consistency across their communications,” Parry-Jones said.
Some of DeepL’s customers are Japan’s Nikkei, US software company Zendesk, and Deutsche Bahn, Europe’s largest railway operator and infrastructure owner.
“Deutsche Bahn relies on DeepL to connect its multinational employee base of thousands across multiple markets. Given that only 64% of employees are based in Germany, effective communication across multiple languages is vital for the company’s international success,” Parry-Jones remarked.
Moreover, Deutsche Bahn’s employees often work with highly technical and industry-specific terminology, requiring standardised translations to prevent miscommunication across departments and roles.
“Deutsche Bahn has improved its internal communication processes with DeepL’s translation tools, including its glossary functionality, which helps standardise terminology across teams. Recently, the company introduced an AI-powered glossary generator that allows businesses to create custom glossaries by uploading files. This streamlines the process in comparison to traditional manual translations or find-and-replace methods,” Parry-Jones said.
Solid foundation
Earlier this year, DeepL introduced its next-generation LLM designed specifically for language applications.
“This allows for more sophisticated and precise translations and writing for a range of applications, while also reducing the risk of hallucinations and misinformation,” Parry-Jones shared.
Unlike general-purpose models that primarily rely on publicly available internet data, DeepL incorporates proprietary data sets specifically tuned for translation and content creation. This approach is further enhanced by human-led training, with language experts refining the model to prioritise quality.
“Blind tests demonstrate that DeepL translations are preferred by language experts 1.3 times more often than Google Translate, 1.7 times more often than ChatGPT-4, and 2.3 times more often than Microsoft,” Parry-Jones said. He added that these comparisons highlight tangible differences in translation quality, which DeepL’s customers reportedly notice in their day-to-day operations.
Meanwhile, a 2024 Forrester study further underscored the platform’s impact on business efficiency. It found that organisations leveraging DeepL experienced a 90% reduction in translation time and a 50% reduction in workload. These improvements contributed to a 345% return on investment for businesses, highlighting the potential of translation technologies to streamline operations. According to Parry-Jones, these gains also allow enterprises to expand into new markets with greater speed and scalability.
Privacy and security remain top priorities for DeepL, particularly for enterprise customers handling sensitive information. Parry-Jones noted that all texts processed by DeepL Pro subscribers are deleted immediately after translation to ensure confidentiality.
“Many AI solutions lack sufficient security measures, which has led to a lack of trust among businesses and uncertainty among users when it comes to integrating these tools with documents used in the workplace,” he added.
To address these concerns, DeepL adheres to rigorous enterprise-grade standards, including compliance with GDPR, ISO 27001, and SOC2 Type 2. The platform, asserts Parry-Jones, also uses advanced encryption technologies to secure data throughout the translation process.
Window of opportunities
The rapid development of AI as an enterprise tool makes it difficult to predict exactly where the industry — and DeepL — will be in five years, Parry-Jones acknowledged. However, he highlighted speech translation as a key area of focus.
“We recently launched DeepL Voice, a solution that aims to enable participation from all stakeholders and create new avenues for connecting relevant knowledge to the right projects and people. We provide real-time, reliable speech translations for everyday business communications, including virtual meetings and face-to-face communication,” he said.
Parry-Jones emphasised that the ultimate goal of DeepL Voice is to deliver both speed and confidence in translations. He explained that the tool enables teams to engage in conversations with clarity, fostering better understanding and more decisive actions. By addressing language barriers, DeepL Voice facilitates communication and collaboration across multilingual workforces, with the potential to improve operations.
