Auto-translating training into every location language

A franchise network covers Belgium, France, Germany, and Spain. The frontliners speak Dutch, French, German, Spanish, Polish, Romanian, Arabic, and English depending on which location and which shift. The L&D team has two writers and a budget that does not stretch to twelve translation contracts per course. Auto-translation is what closes that gap — one course built once, automatically rendered in the learner's language, with the brand-standard content unchanged. ## The translation problem at franchise scale The traditional translation workflow is: build a course in the source language, send it to a translation agency, get back twelve language files, import them as separate course versions, maintain twelve versions every time the source changes. This worked when you launched a course twice a year and your network covered three languages. It does not work when you launch courses monthly across eight markets with a dozen frontline languages. The failure modes are predictable. Translations lag the source by weeks. Updates to the English version do not propagate. The Polish-speaking frontliner at site 47 trains on a course version that is six months out of date. Compliance writeups assume "the network has been trained" but actually only the source-language locations have. ## What auto-translation gives you Auto-translation in a training platform means the source course exists once. When a learner opens it, the platform renders it in their preferred language on the fly. Update the source — text, video captions, knowledge-check questions — and the translated rendering updates next time anyone opens it. No separate course versions. No translation queue. This is not the same as Google Translate on a webpage. A real auto-translation system for training: 1. **Translates context-aware.** "Sub" means "sandwich" in a QSR context, not "underling." The system uses your industry vocabulary and terminology consistently. 2. **Preserves your terminology.** Brand terms (the menu item names, your loyalty program name, your service principles) stay in the source form across all languages — "Big Mac" is "Big Mac" in Polish, not a literal translation. 3. **Translates video subtitles, not video audio.** Subtitle generation is fast and cheap; voice cloning is not. For 95% of training, subtitles are enough. 4. **Lets the franchisee correct.** If a translation is wrong or awkward, a local Educator can override the translation for their scope. The override does not affect HQ's source. ## What stays human Auto-translation does not eliminate human review for everything. Two cases need humans: **Compliance and regulatory text.** If your jurisdiction requires harassment-prevention training language meet a specific standard, a legal review of the translated version is non-negotiable. Auto-translate the draft, then have legal review the regulated language before publishing. **Brand-voice critical content.** Marketing-adjacent training (the customer-experience script, the onboarding welcome) carries brand voice. Auto-translation gets you 90% of the way; a brand-aware human polishes the last 10%. For the bulk of operational training — SOP rollouts, product knowledge, system walkthroughs — auto-translation is fully sufficient. ## What the frontliner experiences The frontliner at site 47 in Antwerp opens a course assignment on their phone. The platform recognizes their preferred language (set on their profile, defaulting to the location's primary language). The course opens in Dutch. Modules, quizzes, video subtitles — all Dutch. They never see a language picker, they never know there is a French source, the course just works in their language. If HQ updates the source course tomorrow, the next time the frontliner opens it, they see the updated Dutch version. Same workflow, no new file, no translation lag. ## What this enables Networks that ran translation through agencies typically launched courses 2–4 times a year, with translation as the bottleneck. Networks running auto-translation launch monthly. The bottleneck moves from translation to authoring — which is the right place for it. Aristotl's auto-translation works at the rendering layer for this reason: the L&D team writes once in the source language, the platform handles delivery in every learner's language, and brand standards stay consistent across the network. Educators at the franchisee level can review and override specific translations within their scope without affecting HQ's source. ## Per-language defaults vs per-learner preference Two configuration patterns work, depending on your network: 1. **Per-location language.** Each location has a primary language (set by the franchisee). All frontliners at that location see training in that language by default. 2. **Per-learner language.** Each frontliner sets their own preferred language on their profile. Useful in markets with mixed-language workforces (Brussels, Montreal, Miami). Most franchises start with per-location defaults and let frontliners override on their profile. This handles the 80% case (location language) cleanly while not forcing a non-native speaker into a language they cannot follow.

Ready to put this into practice?