Work / Efficient Localisation
Localising a course and producing it again are different jobs. Most workflows can't tell them apart.
Traditional e-learning localisation means rebuilding: recreating each slide, re-typing on-screen text, re-recording and re-embedding audio, re-inserting images into a new deck, for every target language. The work duplicates, the typo risk multiplies, and a course with ten video nuggets across several lessons does not scale to a fifth language. Efficient Localisation exists because that arithmetic is bad.
At a glance
Text, audio and images live outside the deck — text in a spreadsheet, audio and images in folders — and are linked into it by a tool we built for the purpose. The deck holds the structure: layout, timing, animation. Those are the parts a translation doesn't change.
On-screen text sits in a spreadsheet column mapped to a slide number. Edit the cell, refresh the link, and the slide updates. Nobody re-types anything into a text box, which is where localisation typos are actually born.
Each line of narration has a defined naming convention. Record in the target language, save under the same name, and it links to the right slide automatically. Images work the same way — replace the file, reopen, done.
Interface language reused across every course — progress, resources, transcript, standard messages — lives in a global sheet. Translate it once and every course in the library inherits it consistently.
German runs long, Japanese runs short, and neither respects the text box the English was designed for. Text box resizing remains a manual, slide-by-slide pass. It is the step we have not automated away, and we would rather say so than imply otherwise.
The link refresh is a real step. Skipped or saved in the wrong order, text can revert — so the workflow has checkpoints, because a method that works only when everyone remembers is not a method.
Full-HD exports are large and SCORM packages have limits. Export-then-compress is a mandatory stage, not an optimisation, and it is built into the run rather than discovered at delivery.
This is how slide-based courses re-version. Video generated in Scenova takes a different route — a second language is simply a second generation — but the principle is the same either way: build once, and make the language a property of the output rather than a reason to start again.
More on AI Transformation →Tell us the shape of it and we will tell you how we would run it — including if we think someone else is the better fit.
