Professional translation has changed more in the last 24months than in the previous 24years. What was once the exclusive domain of human translators can now be done to a very high standard using AI — in some cases. In others, machine translation is still clearly insufficient.
When AI is enough
Internal documentation: manuals, procedures, internal communications. Accuracy matters but cultural nuance does not.
Digital products and software: interface text, error messages, app copy. Specialist translators still outperform AI for user-facing product content, but AI is sufficient in many cases.
Bulk catalogue content: product descriptions in e-commerce with thousands of items. The cost/benefit of human translation is not justified; AI with selective human review is the right option.
B2B technical communications: emails, proposals, technical documentation. Technical accuracy matters more than stylistic elegance.
When you need a human
Client-facing brand content: corporate website, advertising materials, creative copy. Cultural nuance and brand voice require a human translator who understands both cultures.
Marketing and advertising: metaphors, wordplay, and emotional nuances that work in one language rarely work literally in another. Transcreation, not translation.
Legal documents: contracts, terms and conditions, regulatory documentation. The cost of a translation error is enormous. Always use a specialist human translator.
Sensitive or cultural content: medicine, mental health, religion, politics. Cultural nuances are critical.
The stack we recommend
DeepL for quick, good-quality translation. Better than Google Translate for most language pairs. Claude or GPT-4 for contextual localisation (where you need it to understand the target audience, not just the language). A specialist human translator for everything that goes out to clients under the company’s brand.