This article was originally published in June 2019 and has been updated.
Many companies already use instantaneous translation engines—Machine Translation (MT)—for technical documentation and large volumes of non-critical, external material. However, MT can now be effectively applied as part of an integrated human and MT workflow to improve many of an organization’s language requirements. Great examples of MT application include email, real-time chat, user documentation, and even digital marketing content in cases where the MT engine has been trained effectively and integrated with more traditional human translation processes.
A compelling application of MT is its ability to support global communication. In a recent outcome, ULG worked with a live chat platform that enabled an organization’s workforce to communicate across 14 languages in real time. We trained the chat engine on key terms and phrases that specifically encoded the focus area of the organization and improved the voice, tone, and style of the MT output to be more in line with the nature of the internal discussion taking place. This enabled users worldwide to discuss important internal issues in real time without language barriers and aligned internal views of reports and data so that there were fewer gaps in information or misunderstandings across regions that had previously hampered overall communication efficiency.
Understanding the differences between off-the-shelf vs. customized solutions
When internal employees want to use MT in real-time communication, they will often turn to the web, to a service like Google Translate, which is used daily by millions and offers helpful ad-hoc translations. Where the differences and considerations for an organization lie, are in the security and overall voice, tone, and style of the required output.
From a security perspective, any content input into a publicly and freely available service naturally flows into the public domain, meaning that it can be tagged for search and recall. If a piece of internal communication were to be copied and pasted into Google Translate, it would then become searchable by the wider internet. Obviously, for new product announcements, or sensitive internal data relating to mergers or corporate strategic development, this could pose a significant risk and concern.
Similarly, off-the-shelf public engines are geared towards full-coverage global content, and the output is quite general in nature. As your content becomes more specific – with key terms and phrases relating to your individual industry – or where you have high-value brand or content engagement requirements, the output becomes increasingly less likely to hit the mark.
A customized solution from a provider like ULG addresses both the security and the voice, tone, and style challenges presented by off-the-shelf or public domain solutions. Each MT solution provided by ULG receives a custom and secure token that ensures no one else has access to the engine; it also means that all training and development of the engine is not only specific to the MT engine, but that the MT engine training is not confused by other external content coming in with “alternative” voice, tone, and style input. The outcome for the organization is a reduction in risk of internal communications being leaked prematurely to market and a significant improvement in voice, tone, and style of the MT output.
Connect your corporate world
Using MT as part of a comprehensive global communication strategy can create significant benefits that improve the flow of data across regions and allow both internal and external resources to better understand and connect with your messaging. To maximize benefits while controlling risk, consider your organization’s data security requirements, as well as the value created by ensuring that the established voice, tone, and style are preserved in the final output.
If you are considering integrating Machine Translation into your communication strategy, we encourage you to check out our platform OctaveMT and speak directly with our industry-leading data scientists. We look forward to connecting with you!