case work
CASE001
Building AI Translation for a Global Organisation
AI translation capability that amplified human intelligence
CHALLENGE
A multinational consumer goods company with operations across 165 markets had a critical scaling problem. Their content translation and approval process was taking four to six months per product launch, creating significant bottlenecks for international expansion. Despite strong market presence and growing demand, they needed to move faster. New, custom-built technology with AI translation capability was required.
INSIGHT
The solution couldn't just be faster. It had to be smarter about where to use human vs machine expertise in the translation process. Sales managers, international partners, creative teams, and product teams each carried specialised knowledge about their markets that no system could replicate. The architecture had to amplify that intelligence, not replace it.
ARCHITECTURE
We began with structured stakeholder interviews and quantitative analysis to validate the problem and scope. Evidence-based personas mapped the real pain points across every role. From there, a Product Requirements Document, 100+ wireframes and a complete UX/UI design system were developed and presented. The translation model itself was designed as a human and machine partnership: AI generates the first draft, specialist human reviewers apply the market knowledge, cultural nuance, and contextual judgement that no model can replicate. We worked directly alongside engineering to ensure technical viability before a line of code was written. Twenty-plus stakeholders were brought together to prioritise features, agree a phased rollout, and align on success metrics.
CHANGE
A development-ready product foundation for engineering, with validated user needs, solid technical architecture and full organisational buy-in. A process that had taken four to six months, redesigned to move at the speed the business actually needed. AI native to the operation. Human expertise for judgement.