Translating with Articulate Localization
Articulate Localization is a new part of the Articulate 360 suite of eLearning development tools that allows for automated translation and translation review of Rise 360 and Storyline content, into multiple languages. It even includes the option to choose the level of preferred formality.
This is a big step forward in terms of making it easy to localize eLearning content (truly push-button), but it is still machine translation, without any human involvement, so the use cases really need to be considered carefully.
We’ve had the opportunity to demo Articulate Localization and we’re going to share the ins and outs to help you decide if it might be a good fit for your translation needs.
What are the "Ins" of translating with Articulate Localization?
Articulate Localization has a lot of great features. The primary benefit is that it provides for near-instant translation into 70+ languages by just a few clicks.
Simply go to the translation option, select your language(s), the optional level of formality, and click “Translate”. Within seconds your course content will be translated into your selected languages.
Once your translation is complete you can move to the review functionality that will allow for a human reviewer to take a pass through the translations, making any preferred changes to the translations. Once your review is complete, and assuming you’ve dealt with any images, audio, or video assets that need to be localized, your translations are ready to go live.
Benefits of Articulate Localization:
Automated translation directly in the platform
Support for 70+ languages including right-to-left
Quick turnaround time
Easy to use interface
Built-in review platform
What are the “Outs” of using the Articulate Localization platform?
As I get into the “Outs”, I’d really like to point out that I’m intentionally avoiding making this a “Pros/Cons” breakdown of the Articulate Localization platform. The reason for that is that the translation workflow that the Localization platform provides certainly has its place for certain project types, languages, and quality goals. Here are some points to consider when translation is a need that will help you decide whether the Articulate Localization platform is a good fit for your project, or if you require an alternate option:
Considerations when using the Articulate Localization Platform for your Translation Workflow
It is Machine translation
The translation provided by Articulate’s Localization platform is machine translation. This is fully automated translation performed by a machine, with no human involved. As I said above, this isn’t a knock against the platform, this is just an alternate workflow and level of output that needs to be matched to the project. If you’re translating very general content (with very limited concerns about the quality of the output) into a very well supported language, like Spanish, automated translation may be a good fit. If you’re translating sensitive content, or have any concern over the quality of the translation, for example safety training or anything life science related, automated translation is likely not the best fit.
While there is translation review functionality, it is limited, and as I’ll get into next, there are many additional things that a human-in-the-loop, or more complex workflow, will bring to the table to ensure you maintain the quality of your translation (while still managing cost and timing).
No way to incorporate translation assets
One of the most important components of getting a good translation is the incorporation of translation assets. This includes things like a translation style guide, glossaries, and most importantly, translation memory. These are all assets used by a human translator when performing translation to match the translation to the customer’s specific needs. A great example of where this is important is when the client would like to participate in a review of the translations themselves (in-country review). This is a highly valuable QA step that can add client preferences around specific terminology, branding, or even country specific detail about a product or service. It’s important to incorporate and capture that feedback for future translations. This would require working outside of the platform and using the above mentioned assets to capture the client specific feedback.
This is what an alternate workflow provides. By exporting the content for translation, bringing these assets to bear, and then re-importing the final translation back into the platform. This is not an available option within the Localization platform currently.
Limited review functionality
Building off of the above two points it is easy to see why a review of the automated translation would be critical. While the Localization platform does have review functionality built in, it is limited, and again may not match the output requirements of most translation projects. Features like change tracking, commenting, and user and audit tracking / reporting, are all important components of the translation review process. This is especially true in a workflow represented by the Localization platform where the review step involves a human translator reviewing machine translation for the first time, which may require a significant amount of modification. There are translation companies that provide their own platform for translation review, which again involves the standard workflow of exporting and re-importing translations.
Limitation in the Application of Machine Translation
As mentioned previously, we are big proponents of the application of machine translation, and machine / AI enabled translation workflows. That being said, we’ve found that to get the best output out of a automated translation workflow you need to use the best tool for the job. For us that means constantly researching and testing all of the current machine translation engines (Google Translate, AWS, DeepL) and LLMs (ChatGPT, TAUS, now DeepThink), and assessing which engines create the best output. This will vary depending on the content type, the complexity of the content, and the languages required. The optimal workflow may split translations up into multiple routes depending on the requirements. The Localization platform currently relies only on one translation engine and does not have the ability to differentiate based on any of these factors. Again, this is not necessarily a knock, because it does depend on your specific quality requirements, but it is certainly a consideration.
First Look at Articulate Localization Platform - Conclusion
We were very impressed with the functionality of Articulate’s Localization Platform. For the right user, and the right project type, this can be a powerful tool for enabling multilingual content and providing access to non-English speaking learners. I’m sure that as development moves forward with the platform there will continue to be upgrades that will help address some of the “Outs” we described above.
If a high quality translation is required, or if there are any questions or concerns about the quality of a machine translation and whether that will match your specific project requirements, it’s probably still best to consult with a professional translator or translation company to make sure you’re getting a workflow that matches your needs.
And of course, as always, if you’re interested in learning more please feel free to contact us for a chat!