Full-Store AI Translation System with Brand Voice Rules

2025-06-01

Products, collections, pages, menus, and SEO. Translated at scale with brand voice preserved.

Built an automated translation system for a full Shopify store. Products, collection descriptions, pages, menus, Klaviyo email content, and SEO meta all translated into multiple languages. Brand tone rules and SEO structure enforced throughout.

The challenge

Expanding an ecommerce store into new language markets means translating thousands of strings: product titles, descriptions, feature bullets, SEO meta, collection copy, navigation menus, pages, and email flows. Machine translation produces generic output that breaks brand voice. Manual translation is too slow and expensive at catalog scale. The goal was a system that translates at speed but preserves tone, SEO intent, and brand rules.

Approach

• Built translation rules for each target language: tone of voice, forbidden phrases, SEO keyword handling, and brand-specific terms
• Product translation automated via Shopify's Translations API: titles, descriptions, feature bullets, and SEO meta all translated per product in batch operations
• Collection copy, page content, and navigation menus translated via the same API layer
• Brand voice enforcement built into every translation prompt: outputs are reviewed against tone rules before being written to the store
• SEO structure preserved: translated meta titles and descriptions maintain keyword intent for each target market
• Klaviyo email flows translated separately: email copy adapted per language with matching brand voice
• Verification scripts built to audit coverage: which products have translations, which are missing, which need review

Results

• Multi-language — full store translated: products, pages, menus, emails
• Brand voice — rules enforced in every language
• SEO-aware — translations preserve keyword intent per market
• Automated — audit: coverage, missing translations, review flags

Tools used

Claude AI, Shopify Translations API, Klaviyo API, Python, Prompt Engineering