Advanced translation technology is like any other tool: it can be the solution or the problem. When it’s the problem, it can create issues that filter throughout your content and user experience, damaging your brand’s reputation and cutting into your bottom line. However, when it’s the solution, you can save thousands—or even millions—while significantly expanding your reach.
But what decides whether your advanced translation technology will be the problem or the solution? The core deciding factors come from how it is initially set up and how you use it. If you leverage it correctly, you can easily streamline tasks and simplify your localization projects. This is especially true when using some of the more popular translation technologies available on the market, like machine translation (MT), computer-aided translation (CAT) tools, and artificial intelligence (AI).
Tips for Using Advanced Translation Technology
Success with your advanced translation tools will all depend on your use and maintenance expectations, so it’s vital to go into any tech-enabled localization project with definite goals in mind. That’s especially true of the three most popular technologies: MT, CAT tools, and AI.
Machine Translation (MT)
Machine translation—like Google’s Neural Network or Microsoft’s Translator Hub—allows you to quickly translate content into many different languages. While it is a useful tool, it’s essential to not rely on it alone. Most companies are unaware of this and assume it’s an excellent way to save money on translations. They’ll use it to translate their content roughly and then send it off to linguists for a quick review. This process creates major accountability and trust issues, which filters into the content.
MT works far better for two specific use cases:
- Testing: Machine translation can act as an early warning that your app’s coding isn’t up to par. You can use MT to translate and process content in-app to see if the code will remain intact during the transfer process. This strategy can help you identify pertinent issues, like text boxes that aren’t large enough or section breaks that don’t fit accurately. With early MT tests, you’ll be able to ascertain exactly how much code cleanup you’ll need to perform in order to move on to the actual translation project later.
- Gauging language demand: If you have pages with a low view count that you wish to publish in multiple languages, MT can help you get it done quickly. You can post the MT version of these pages and then use traffic and user reports to gauge interest. If, for example, you push an MT page live in Finnish and see a sudden uptick in views—as well as user requests for support—you know this is an area of high demand that you can target for a full-court translation. When doing this, it’s vital to put a disclaimer somewhere in the content that lets readers know it’s machine translated.
Computer-Aided Translation (CAT) tools
CAT tools allow translators to control their workflow. They provide information linguists need to complete better translations. CAT tools include MT, corporate lexicons, glossaries, and spell checkers. These can vary widely from platform to platform, but they all have one common goal: to help linguists translate strings faster while maintaining high accuracy.
One common flaw among many CAT tools is that most are entirely abstract from the source. The linguists will see only strings of content to translate; they won’t see, for example, the website or UI where the content will live. This creates an issue since much of the translation is contextless. Given this, it’s essential to explain the context to the translator with the help of these tools.
Artificial Intelligence (AI)
AI is a hot topic right now, and the localization industry is no exception. The ability to automate tasks based on human-like insight and computer reasoning is considered an ideal way to improve translation tasks for everything from workflow to hiring.
AI is something best to approach on a case-by-case basis. You have to carefully review the automation potential of all your tasks, and specifically, what rules can be inferred from those tasks. It’s about building a system that’s rule-based rather than exception-based. As a result, AI is something best used for customized functions for a particular goal.
Making Translation Tools Better
The main way to make advanced translation tools perform as (or better than) expected is to focus on the end-to-end solution, potentially through a localization management platform. You don’t want to solve problems in bits. You want to create one solution that filters through everything.
Translation memories (TMs) are another important tool to help round out the rest. Critical for translators’ efficiency and quality, TMs constantly evolve with each approved translation project and inform the projects that follow. A translation memory, along with glossaries and style guides, becomes a critical part of the productive CAT tool environment. The ideal scenario for the most effective localization is that linguists are able to understand the context as well as the basic strings needing translation. Developing this context is more than just an automatic suggestion by MT or AI; sophisticated localization calls you to take your global operation to the next level.
Advanced translation technology is not a turnkey solution to localization management—instead, it’s a way for you to improve your quality and efficiency while completing a project. By using the various options for the right reasons, you can expedite your process and keep costs down.
September 3, 2020