Need for Speed and Quality? Unbabel’s Translation Services are Here

February 18, 2022

When looking to enhance your organization’s multilingual communication capabilities, you want a solution that produces high-quality language translations, fast. In researching your options you may have noticed that human-only translation services guarantee accuracy but are often costly and difficult to scale. Solutions that rely 100% on AI technology, on the other hand, provide timely translations but the quality can vary widely. 

Is it possible to get the best of both worlds? With Unbabel, you can. Read on to learn more about how we can help your business strike the right balance between agile customer service and high translation quality in multilingual communications.

In customer service, speed is non-negotiable 

In our 2021 Global Multilingual CX Report, a survey of more than 2,750 consumers in the US, UK, Brazil, France, Germany, Japan, we learned that customers want their questions answered and problems resolved promptly. When we asked respondents about their top three priorities regarding native-language customer support, 47% listed quick response time and 46% said fast resolution times. Clearly your global customers have a need for speed in customer service. Unbabel can fulfill that need with our AI-powered, human-refined language translation services. 

Supercharging translation efficiency 

AI is the driving force behind our ability to produce native-quality translations in a fraction of the time that human-only methods would take. By supplementing human skills with advanced machine translation technology, the Unbabel Language Operations (LangOps) platform allows organizations to drastically improve speed and scalability without compromising on quality. The average human translator can process approximately 300 to 400 words per hour. With Unbabel, that person can translate 1,200 to 1,400 words per hour. 

This boost in translation efficiency can help your organization provide faster multilingual customer support and increase customer satisfaction. For example, the location technology pioneers at TomTom used Unbabel to achieve a 42% reduction in overall email turnaround time as well as a 10% decrease in chat abandonment rate. 

Striking the right balance 

Customers’ expectations for timely communications depend largely on the channel they are using. That’s why the ratio of human-to-machine translation executed by Unbabel is custom-tailored to the situation and complexity of the request involved. 

Because customers anticipate instant responses in live chat, these messages go through our augmented machine translation process and are delivered in seconds. To ensure the quality of these translations, we regularly select chat messages to be reviewed by our community of post-editors, and any edits are used to train the company’s translation engine and improve its ability to produce top-notch translations autonomously. 

With a channel like email, where customers understand they aren’t going to get an immediate reply and each response will likely be longer and require more nuance, we send every machine translation to our community of editors for a thorough quality check. For static communications, such as FAQ pages and marketing content, premium quality is essential and turnaround time is less of a concern. These machine translations are quality-checked by multiple editors and then given a final review and sign off by an accredited senior community editor.

Quantifying quality across translations 

At Unbabel, we are proud to be on the forefront of assessing and improving quality in tech-enabled language translations. Our research team is filled with many of the world’s preeminent scientists, engineers, and scholars on language technology and translation quality. In just the past few years we have made significant contributions to the translational research community, including the COMET framework/metric and MT-Telescope, a tool for understanding the performance of machine translation systems. Let’s take a look at some of the factors that allow us to give global organizations the high-quality translations their customers deserve. 

Brand-specific translation engines 

We use a process called domain adaptation to train our customers’ machine translation engines with historical data that reflects their brand’s unique terminology, style, and tone. If a company needs translations across several different language pairs, each pair gets its own fully-optimized translation engine featuring translation memories, glossaries, and style guides to allow the brand’s voice to be consistently captured in every language. 

Unbabel also runs automatic retaining processes to keep machine translation engines up to date. This involves taking new data from a customer and adding it to the old data, using that combination to retrain the model, evaluating performance, and potentially deploying a new model if necessary. This process of monitoring and retraining facilitates continuous improvement and ensures that brand standards are always being met. 

Sourcing high-skilled editors 

We have a special system in place to guarantee that our community of editors includes only expert linguists. Each translation handled by an editor has a rating assigned and the overall average makes up their rating score. This process helps us determine that our editors are producing excellent work that will be perceived as native quality by the target audience. If an editor’s rating slips below a certain threshold, they are demoted to trainee status and they aren’t able to access Unbabel paid translations. 

LangOps: The key to superior business translations 

Unbabel’s LangOps platform provides a powerful combination of speed and quality that other professional translation services just can’t match. It also supports a unified language strategy across all departments of your organization, compared to the siloed approach that goes along with traditional localization. That’s why over 90% of decision-makers we surveyed agree that LangOps could be extremely or very important for their organization’s global market strategy. 

Still have questions about how this technology works and why it’s the best way to produce multilingual communications? Schedule a demo and we’ll show you how far your business can go with Unbabel.

About the Author

Profile Photo of Phill Brougham
Phill Brougham

Director of Product Marketing at Unbabel, Phill Brougham spent the last five years working for SaaS businesses focused on applying artificial intelligence to solving real-world business and productivity problems. Throughout his roles, Phill’s focus has been on translating technological capability into clear, understandable value.