Machine translation (MT) and neural machine translation (Neural MT - with embedded knowledge) is especially well-positioned to address the growing volumes of localized translation and localization due to its inherent scalability and reduced price.
Everyone can use this technology. At least basic users and those who don’t have very demanding quality requirements.
But preparing a whole MT or NMT system capable of solving the translation needs of a company is not an immediate task.
Companies that choose to use these systems have some technical work ahead of them.
Ideally, you need to have the translations - your own documents or others that can be used as a base - that, when properly prepared and included in the systems, will allow the creation of a properly trained engine.
These systems can still "learn" continuously and in real-time.
Through the process of correcting sentences that leave the system and then, once corrected, re-enter the system, the system can learn and improve the next results that are returned when found in similar sentences (blocks).
This "intelligence" (AI - Artificial Intelligence) permits the creation of what we normally call Neural Machine Translation.
HCR uses these systems on a daily basis and can help its clients to develop similar engines to translate their company's content.