As the demand for quality affordable translations is ever increasing, the technology to deliver better and faster results is continually being improved upon. Companies are moving beyond traditional translation memory and terminology management systems to find new ways to leverage automatic technology-human hybrids.
When using artificial intelligence (AI) for translations, its role is often misunderstood. While playing an important role in today’s machine translation technologies, it also has significant applications in many other aspects of the language translation and localization process. These applications include AI powered file analysis, estimating turnaround times, language resource management, translation provisioning, translation reuse, linguistic review, and risk management. Let’s take a quick look at some of the ways Stepes is using AI to help streamline parts of the translation process:
File analysis is the process of extracting text into a format that is easily useable for linguists to work with. This analysis is used to calculate word count and cost estimates among other things. Traditionally file analysis has been performed by a team using CAT tools to get those word counts, and to determine what if any fuzzy matches they get from translation memory. This manual process had a multitude of human touch points which on top of being time consuming, runs the risk of misplacement as it is shuffled and sent between multiple resources.
“Stepes AI powered translation management system automates the file analysis tasks by automatically detecting the source text language, leveraging TMs on the cloud, and providing translation quotes in real time. In addition to regular file types, Stepes has expanded our AI capabilities to include scanned documents and image files which extract text in real time, on-demand. This highly automated process is a game changer because Stepes customers no longer have to wait for 24 hours or longer just to receive a translation estimate. In today’s digital economy, a single day’s delay can cause lost business opportunities with millions of dollars in lost revenue.”
Assigning translators to actual projects is an integral step in the translation process. Once a project manager identifies a translator for a particular project, they must then send files needing translation and other relevant information to the linguist. The vast majority of localization industry today works with freelance translators who are based in-country, meaning the potential for production bottlenecks as email messages and file transfers are organized and sent.
Stepes has overcome these hurdles head on by creating an automated system of sending the right projects to the most qualified translator for the job. They are able to work within the platform to access all required material in real time.
An interesting challenge that language service providers face is the ability to achieve fast turnaround times when there is a great fluctuation in the amount of work coming in. In this fast-paced digital economy, we are seeing increasingly higher demands for content to be quickly deployed, however rushing a translation project often has financial drawbacks.
What Stepes has done to meet this challenge is to create a AI enabled project management workflow. It uses distributed project assignment APIs to perform tasks such as splitting segments into multiple jobs to be performed simultaneously by different translators, automatically selects qualified linguists and pushes notifications to them, and merges individually translated segments to create a seamless final document.
We have seen a lot of advancements in translation technology over the last decade. With an ever growing globally connected world, communication is key. Digital content is especially fast moving and global enterprises must be ready to deploy up-to-date content on a continuous basis. As the demands for quality, speed, and accessibility keep increasing, so too will the technology that supports it.