European language service and technology provider Tilde successfully implemented the XTRF Translation Management System in record speed. In fact, it took them less than six months to move all their operations across from their old, internally developed, and highly customized project management system, to XTRF.
They’d laid the groundwork
Sure, it may sound pretty incredible, but this super-fast transition was a testament to the work of the XTRF team. Their dedication and experience played a huge part in facilitating the processes, as did their openness to customization and external app integration.
What’s more, the on-site and online training sessions provided by XTRF also enabled Tilde’s employees to quickly get to grips with the new system, thus ensuring efficiency from the get-go.
Perfect doesn’t exist
There is no one-size-fits-all in translation management automation when you are dealing with complex and challenging projects and innovate by bringing new technologies into it. That’s why the ability to customize the XTRF system was vital for Tilde when weighing up their project management system options.
That said, they’ve already seen a big difference in certain areas thanks to certain customizations. For starters, the macro and virtual columns which generate reports with every job delivery enable them to keep better track of their vendors.
For example, if there is a vendor who is repeatedly submitting projects late, the system will pick up on this, giving the company a chance to resolve the issue or talk through any potential problems with the vendor.
The use of macros for invoicing has also helped the company cut down on time previously dedicated to administration tasks, and to better keep track of the finances in general.
It is estimated that Tilde will process about 70 000 incoming translation tasks with XTRF in 2021.
Friend or foe?
Tilde has long been using Machine Translation, even back in the days of Statistical Machine Translation. They’re an innovator in this area and their in-house translators are the lucky first to benefit from the new technologies they implement. Now, Tilde’s MT Dynamic Learning engines learn from translators’ corrections in real-time, adapting the “voice and style” they are using to produce more fluent and accurate translations.
As their translators explain, “ Dynamic Learning opens a new dimension in the Neural Machine Translation landscape. It analyses every correction I make and consequently, the translations require far less translator engagement as they are more human-like, and I can translate greater volumes in less time. ”
Indeed, elsewhere in the industry, those who once viewed machine translation as a threat, now better understand how it can be a valuable tool for increasing productivity, translation volumes, and ultimately, revenue.
XTRF in cooperation with Tilde localization team has added Tilde MT to the project management process by creating a direct Memsource integration workflow for those client projects where MT can be used. This enables their translators to increase their productivity and keep up with increasing translation volumes without the extra stress.
There are now over 40 Tilde MT language pairs. Often translators choose to use custom MT systems that are adapted to their specific domain and language. And now they can select a dynamic learning engine, one engine which ensures simultaneous learning and adjustment for multiple projects and specific domains.