Ahua

Act I — Inception

Why does a fertilizer company need a design system?

That was year one of Yara's digital farming. A century-old fertilizer company on a mission to feed the world had just opened hubs in Berlin, San Francisco, São Paulo and Singapore to build digital products for farmers. A dozen teams trying to solve different parts of the same puzzle, without a shared product language, and the differences costing real money in duplicated work.

The designers saw it first. By 2019 two hubs had started design systems of their own: Fertilize in Berlin, Ahua in Singapore. By 2020 the design system I had started in Singapore had become the foundation of Yara's global standard.

2018–2025

One Yara, one language

Act II — Earning it

How does something nobody must use get used by everyone?

In 2021 Yara merged its digital hubs into one organisation and needed a single design system. Ahua was already there, built on field research and proven across markets. Yara made it official.

A complex system that works is invariably found to have evolved from a simple system that worked.

Official, at Yara, means less than it sounds. The company runs federated: no part of it can require another to use its tools. Going global gave me the title and not one lever to make anybody adopt anything. Ahua earned adoption by making product teams' lives easier, and it had to keep earning it every quarter after that.

What carried it was a vision for how one system could work across a global organisation, and a process for building it.

Yara's regions didn't need the same thing. In Asia and Africa, teams ship to smallholder farmers over monsoon networks. In Europe, to enterprise buyers on broadband, inside a regulatory regime with its own logic. I turned that difference into the architecture: a system of systems, where regional design systems inherit from a global core and extend it for their markets. We helped Yara Europe establish its own regional system and launch Yara Plus and the next generation of AtFarm. Each region carried the standard forward in its own way.

Getting there meant resolving what already existed. Fertilize was a React library built for the professional market. Ahua was React Native, built for low-end devices. Each had a stack, a history and people who had built it. The harder work was making one team out of two. Bringing everyone behind a single vision took six months.

The process was what teams actually felt. I adopted Zendesk's Triple Diamond and added a fourth by splitting the third into Adoption and Impact: getting a component used and proving it had been worth building need different evidence and different rhythms. Shipping cross-platform with several teams at once needed a different delivery model too, so we moved to Large-Scale Scrum (LeSS): a 2.5-week sprint inside a three-week cycle, with a distributed team of fifteen across five time zones.

If you build it, they won't come. A design system is a fight against entropy, and we had to be the most horizontally connected team in the company, constantly making the case for Ahua. To help product teams adopt it, I worked with regional design leads to set up a support team in each region, embedding design system specialists inside the product teams there. The regional model became the strongest driver of adoption as the product portfolio grew — the system travelled because people carried it and cared for it.

Working with product teams showed us where Yara had no design at all, so we built the Rapid Design Team: designers and researchers who took a request and shipped it in weeks using the GV Design Sprint framework. Sustainability dashboards, an agronomy tool, a finance dashboard for a department that had no design team of its own. One of the proudest moments was Tankmix, a core piece of Yara technology whose creators had left the company and which no team had managed to modernise in years. We rebuilt it, and the system's most vocal opponents became its biggest supporters.

Seven years is a lot of work to fit on one page. A global illustration system, a localised content framework, rural UX guidelines, a design system portal, the Figma enterprise migration, GDPR and European Accessibility Act readiness, cultural intelligence training. Ahua brought the same discipline and the same standard to every level it worked at, from pixels to platforms.

In production

• 
• 
• 
• 
• 
• 
• 

Act III — It holds

What was it worth, and what did it cost?

Six million dollars estimated return over three years. Fifty-plus teams shipping twenty-five-plus products by 2025. Forty percent reduction in design-to-development time.

Time savings, quality improvements and collaboration benefits, over the three years we measured it.

In 2023 Yara transitioned to a fully regional model, dissolving the global disciplines, design among them. To bring Ahua to formal status I took it through the Capital Value Process, Yara's method for justifying major investment. It came out with the support the system needed to be funded.

The hardest thing in seven years was never a component. AI can write code faster than a person can review it. Agents will join the collaboration next. Neither will make fifty teams agree on a shared product language, or help them understand each other, or bring them to work toward a common goal. A design system is a people system.

Ahua is still in production at Yara, used and supported by teams across the company. It outlived the team and the lead.