A national food-security platform
Turning a stalled MVP into a product people actually used — an AI decision platform for managing a country's strategic food reserves.
Client work, shown abstractly. The client is a sovereign-fund company and the platform handles strategic-reserve data, so there are no real screens, numbers, or names here — only the problem, the approach, and what I did.
A government food company needed to see threats to the country’s food supply before they became shortages — global market shifts, shipping disruptions, inventory risk — in one place, fast enough to act on. An earlier version of the platform existed, but the people it was built for weren’t using it.
The real problem
The brief looked like “make the dashboard better.” The actual problem was adoption: the first build didn’t fit how decision-makers worked, so they fell back to fragmented sources and spreadsheets. No amount of visual polish fixes that.
So I started with research the team wasn’t used to doing — interview guides, conversations with the food-security analysts and the executives above them — and found the pattern: a tool only gets used when it draws on data people trust, fits a decision they already have to make, and is easier than the workaround. We rebuilt the platform around those journeys instead of around the data we happened to have.
What I did
- Led UX research and reframed the engagement from “redesign the UI” to “fix adoption.”
- Mapped the real decision journeys — procurement calls, shortage monitoring, market assessment, executive oversight — and structured the platform around them.
- Designed the role-based dashboards (executive, strategic-reserve, trading, operations) and the visual and iconographic system that held them together.
- Shaped a conversational AI assistant so analysts could query the data in plain language — designing the interaction around real questions, not a chatbot bolted onto a dashboard.
- Set the design direction as the platform expanded to new commodities.
Why it matters to me
This is where my work tipped from designer to product builder. The leverage wasn’t in the pixels — it was in naming the right problem, aligning stakeholders who each owned a piece of the truth, and designing an AI feature that fit a real workflow. The platform went live and the team that wouldn’t touch the old one started building their work around the new one.