The use of artificial intelligence (AI) technologies is established in many areas, and most of us are probably using AI every day, at work and at home. Alexa, Cortana and other virtual assistants are just one example. Even Facebook and Google use AI and machine learning to, for example, recognize pictures of our friends in our photos and prioritize email importance.

AI makes it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks. Taking these cues, we can also apply it to learning. This article discusses one of these learning applications: automatic translation.


AI-based translation engines are useful and exciting applications that detect the language of a text string and translate it based on previous translations.

The historical machine learning technique was statistical machine translation (SMT), which would find the best possible translation of a word by grabbing a few words around it to understand the context. In about 2011 came neural machine translation (NMT), which solved a lot of issues with SMT by  translating a word based on the entire sentence instead of a couple of surrounding words.

Consider the inside of the human brain, made up of 7,000 neural networks that process information given to them by our senses. The NMT model is based on the same concept. During the translation process, each sentence is passed through a channel to evaluate each word, enabling a more accurate translation.


Speed and Quality: With NMT, the translation engine can learn linguistic rules on its own from statistical models, results in a well-translated document in seconds.

Cost: NMT may not sound simple, but the process of training and deploying a translation engine is cost-efficient. Additional cost savings come in the quality of translation.

Accuracy: Research has found that when compared to SMT, NMT has much fewer errors.

A Case Study

Proprietary research demonstrates cost reduction of between 30% and 50% and an 80% reduction in production time for global workforces with multiple language needs, including translation of languages that use different characters.

The organization in question is a large, multinational company with employees who speak 29 different languages. Each year, the company rolls out a course to its global sales organization, which takes six months to develop in English. It took an additional six months to translate the course for its Asia Pacific sales employees, which impacted the company’s ability to deliver it in a timely fashion.

With consultation, the company developed a solution using emerging translation technology that shortened the translation time to from six months to one month, saving the organization a significant amount of money in the process. While using a proprietary user interface and the fast processing time of NMT, we have seen time savings of approximately 80%.