In 2020, before the release of NFA, machine translation accounted for 17% of Ariel’s overall translation. NFA’s implementation in 2021 reduced the company’s human translation efforts by an impressive 31%, with the quality of machine translation improving by 100%.
With broader integration into XTM, NFA has the potential to reduce meticulous editing even further by including 85-99% fuzzy matches. By extension, NFA-translated segments have the potential to increase MT translation by 129%.
In addition, having this vastly improved MT engine provides Ariel’s translators with a second, high-quality reference source when translating new segments. While this benefit is hard to measure, it undoubtedly reduces post-editing time.
As Ariel looks to the future, the company is interested to see how they can drive down translation and review costs even lower by focusing on their content creation processes. For instance, the localization/content team is in the process of working with their writers to structure their content using predictable patterns and language. This will enable machine translation to be used to an even greater extent, making localization smoother and more efficient. In addition, the localization team is interested in how they can achieve even greater levels of automation by incorporating AI into all aspects of their translation processes.