← Selected Work
Global FMCG BrandConsumerBuild2023

Marketing automation at scale.

AI content generation and campaign tooling for a global consumer brand — on-brand creative, produced and localised at the speed the market moves.

Campaign engine
Duration
6 months
Team
Pod of 5
Core stack
Next.js · Python · S3
Models
Diffusion + LLM + guardrails
Data
120k brand assets
Reach
40 markets · 18 languages
The challenge

Producing and localising campaign content across dozens of markets was slow and expensive. Brand consistency suffered as volume grew, and local teams waited on a central bottleneck.

Our approach

We built a brand-aware content engine and a campaign workflow around it — generating on-brand variants, localising at scale, and keeping approval firmly in human hands via a guardrail layer trained on the brand system.

What we built

A campaign engine: brief → variants → localisation → approval, with brand guardrails enforced at every step and a review queue for market teams.

Tech stack
Application
TypeScriptNext.jsReact
AI / data
PythonFastAPIPostgreSQLBrand guardrail layer
Models
LLM copy generationDiffusion imageryLocalisation models
Infra
AWS S3 · CloudFrontLambdaTerraform
What we delivered
  • A brand-aware content engine
  • A brief → variant → localisation → approval workflow
  • A guardrail layer enforcing brand & legal rules
  • Market-team review queues across 40 markets
  • Asset library integration (120k+ assets)
Timeline
2 weeks
Discovery & brand-system mapping
10 weeks
MVP for the lead market
6 months
Rollout to 40 markets, 18 languages
Data at work
120k
Brand assets integrated
40
Markets live
18
Languages localised
Outcomes
10×
Creative variants per campaign
Hrs
Localisation, down from weeks
1
Brand system, enforced
Next case →From lab to production — an AI transformation roadmap.

Have something
that needs to ship?

Tell us the problem. We'll tell you honestly what makes sense — even if that's not us.

Start a conversation