Part 3: More Content, Less Cost: Machine-First Translation in the Intelligent Translation Era

Humans are at capacity and unable to scale to provide the amount of content that customers demand. To implement a successful omnimarket strategy, it is critical to look to the intelligent translation era for a machine-first, human-optimized solution.

Translation Memory (TM) has always been the foundation that translation management solutions are built upon. New translation requests are processed by the TM to identify segments with approved translations for each target language. For content without an exact translation, the TM offers similar or “fuzzy” translated content, making the translator’s job faster and easier. By reusing approved translations, the TM ensures translation consistency across markets and document versions while reducing new or incremental translation volumes and costs. 

However, it is worth noting that new technologies are rapidly evolving that may one day soon supplant the supremacy of TM—Neural Machine Translation (NMT) is one such technology. NMT leverages Linguistic AI and machine learning to translate content through a neural network. These automated translation methods, when used with a translation management solution, allow companies to process content for translation much faster and at greater scale than previously possible.

“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”—Buckminster Fuller

Let’s return to our example company, Solar Flare, and examine their support content. The volume of support content in Solar Flare’s knowledge base system (KBS) is too large to be translated on time and on budget with their current process. But with SDL Machine Translation, particularly when integrated into a larger translation management solution, Solar Flare is able to translate all support content into all required languages. In addition, NMT enables on-demand translation for customers speaking languages beyond those currently supported by Solar Flare. And if Solar Flare wants to improve the quality of their best-performing support articles, they have the option of integrating human editing into the NMT process. These NMT improvements result in dramatic increases in user experience and customer satisfaction while lowering support costs and time to market.

By implementing a machine-first, human-optimized solution, meeting the demand for global content becomes an achievable goal. Read our next blog to learn how data can be a game changer for your localization program.