The R&D team developed a unique language and translation workflow solution, integrating previous advancements in AI-generated translations with a user workflow on an open-source content management platform.
The solution is now available to the client, which deals with specialised multi-lingual content, managed by a diverse and geographically dispersed team.
Combine the automatic translation engine with a multi-step workflow, enabling content editors to trigger automatic translations and apply an approval workflow to ensure the accuracy of the translation.
The team was asked to combine the automatic translation engine with a multi-step workflow, enabling content editors to trigger automatic translations and apply an approval workflow to ensure the accuracy of the translation. We achieved this by substantially merging the capabilities of the open-source Umbraco platform and systems like Azure Translation services, using newly developed Application Programming Interfaces (API).
The implementation of the translation workflow enables an organisation's editorial team to provide content in a “master” language, request automated translations for any number of languages; send out notifications to translation professionals; arrange a staged approval process so that translated texts are released in a controlled manner. This is challenging functionality given the specialist language involved, but the team managed to overcome internal complexity and deliver on the project.
The existing website code base required the development team to substantially refactor functionality while also ensuring additional capabilities were delivered. This work aligned with previous work on automated translations and data integrations, and proprietary techniques were used to combine human and machine contributions in a smooth workflow.
Scientific or technological uncertainties
The advancements described above could not be achieved without resolving scientific/technological uncertainties. The main scientific/technological uncertainties that the company sought to overcome during this claim period are summarised below:
- How to material merge the Workflow capabilities of Content Management, automating certain aspects through the integration of a proprietary translation engine that combines information from both systems
- How to perform complex, dynamic, and scalable data exchange between systems, allowing the website to run even if the external system is slow or unavailable
- How to ensure a large, dispersed team of translators can collaborate in a methodical and organised way, yet still make the most of automated translation technology
Eventually, the team was able to circumvent all technological challenges to achieve a seamless user experience, as well as an efficient Translation Workflow platform for the client teams.