3 years ago
November 19, 2018
Autonomous Localization (AL) — The Next Stage in Localization Development
The conventional translation and localization process practiced by most traditional translation agencies typically involves a lot of manual work. Each language involved with every translation project generates overhead and requires email/phone interactions, increasing the cost and turnaround time of the project.
Overhead varies between companies as some agencies use more or less advanced translation management systems (TMS). Regardless, traditional agencies can typically handle up to several thousand projects a month at most—even those who leverage technology are essentially “manual” agencies in their DNA.
One Hour Translation, and other similar online translation agencies, were built from the start as online, internet based operations. At their core, online translation agencies operate more efficiently.
In traditional translation agencies—even those that use sophisticated TMS—the system is a tool designed to assist project and account managers. The account managers run the show, and the TMS is a sidekick.
TMS Takes the Lead
Contrary to that, in One Hour Translation and other online agencies, the TMS—not the account manager—is at the forefront, automatically managing most projects. Account managers or support reps intervene only when the TMS reports there is a problem.
In online agencies, technology drives people and not the other way around. This results in much higher capacity and efficiency.
For example, One Hour Translation manages hundreds of thousands of translation projects a month, all with just a few account managers and support reps. Projects are sent directly to our TMS via API and the system allocates the work, monitors quality and takes care of all other pre-processing or post-processing tasks.
Projects are sent directly to our TMS via API and the system allocates the work, monitors quality and takes care of all other pre-processing or post-processing tasks.
Using TMS to drive project management is not only a boon for agencies, but also for enterprise customers. A TMS-driven approach enables high capacity, direct project launch from one system to another (via API with no manual work required), support for varying project sizes, streamlined processes, and automatic generation of additional tasks before and after the translation (like running TM, DTP etc.)—among other benefits.
Yet despite all of the benefits from what online agencies do, it is just not good enough!
Most business customers see the entire translation process as a “necessary evil”—for them, the ideal scenario would be to click a button and have the localization problem solve itself. After all, this is close to what happens with many other technical tasks: server capacity, backups, etc.
Autonomous, Continuous Localization
Not long ago—before AWS, Azure, and the like—planning and deploying the right servers in the right numbers at the right time was a daunting task. It was time consuming, often more expensive than it should be, and downright irksome.
Today, businesses don’t even think about it. Need more capacity? AWS adds servers on the fly, and you don’t even feel it happen.
This should be the goal with translation and localization: Autonomous, Continuous Localization (AL).
Ideally, customers should be able to activate Autonomous Localization systems with a click. The systems automatically identify new content to be translated, send it to the right type of translation (Machine, Hybrid, Human, Professional etc.), run quality control, post-editing and so forth. Then, the translations are sent back or directly updated in the customer’s system.
No manual work involved, except for human translators if that is the preferred translation type.
This is already happening on a small scale with some localization systems but, in most cases, extensive human intervention is needed for the process to run smoothly. The issue with human involvement is simple: more involvement leads to more errors, issues, etc. A process that requires fewer humans to run will run more smoothly.
Hybrid Translation as Enabler
Hybrid translation (aka PEMT) is a key enabler of the Autonomous Localization vision, mainly because there is less manual work required. When the industry reaches the point where most segments are translated properly by an NMT engine, post-editing or human translation from scratch will become the exception, not the rule. By then, the translation quality and turnaround time will have become more predictable, making way for even more automation of processes, which in turn further decreases the number of “human exceptions.”
Putting all these factors together—with the right connectors to the various enterprise systems—will lead to truly autonomous localization; one that can be activated with a click of a button.
Putting all these factors together—with the right connectors to the various enterprise systems—will lead to truly autonomous localization
Naturally, Autonomous Localization, as well as Hybrid translation, impact the role and work of professional translators and account / project managers. Professional human translators and project managers will still be needed to handle some types of material and certain projects. In general, however, it is safe to assume that within one to three years, hybrid translation and later Autonomous Localization will dramatically reduce the need for manual work in the localization process.
Autonomous Localization is the ideal that customers want. It delivers the most value in terms of better, faster, smoother, easier, and cheaper service. Typically, customers ultimately get what they like.