AI Translations
Cut time-to-publish by 35%, kept content in sync across 100+ languages — with AI output content managers actually preferred.
I established the AI principles the whole product team works from, and applied them to AI Translations and AI-assisted asset management — from discovery through to monetised features. On Translations I A/B tested manual vs. automatic triggers across 200 projects, ran evals on different models before picking one, and designed how AI is monetised in the product.
- Faster to publish35%
- Discovery → MVP8 wks
- Sync · Locale → Source100%
- Of users edited output<1%
- Projects A/B tested200
- Adoption65%
- Stickiness99%




