The challenge
Every campaign had to reach 27 markets at once. Translation went to an external agency, and each launch cost weeks and budget. Product pages were literal translations that missed the words customers in each country actually search for, so they lost visits from Google.
What we did
What people did
With native speakers in each market we wrote a tone of voice and localisation guide: how the brand speaks, which terms it uses and what it never says. We also ran keyword research, country by country, for the 25 product pages.
What AI did
We trained a Gemini model on that guide to localise each campaign's text, PDFs and images. The model took over much of the work that used to go to the translation agency through a connector in AEM, the system where the pages are managed. Native speakers review the output before it goes live.
Product pages for each country
We rewrote the titles, descriptions and metadata of the 25 product detail pages (PDPs) for each market, in line with the running campaign, and adapted the visuals in Figma and Photoshop.
The results
- Campaigns launched in 27 markets, and 25 product pages adapted to each country's SEO.
- Less production time and less translation budget on every campaign.
- The same tone of voice across all markets, checked by native speakers.
What this can do for you
If you sell in several countries and every translation delays deadlines and eats budget, a clear brand guide and an AI model trained on it can do most of the work, with people reviewing. No in-house engineering team was needed.
It’s incredible to finally launch simultaneously with every other market across EMEA without missing a beat. Having perfectly tailored, local-language graphics, disclaimers, and product pages on day one gives us huge confidence when reaching out to customers.