From scattered screens to one coherent logistics app.

How do you navigate and negotiate in Kuyuy?

Kuyuy connects carriers with cargo generators for domestic and international shipments. The migration to a native app promised scalability, time savings, and support — all in one place.

But the reality was messier. The product had many high-fidelity wireframes, but no global architecture, no interactive prototype, and inconsistent terminology across three user profiles (Carrier, Cargo Generator, and SuperAdmin — each with Admin and Employee variants). What made sense internally didn't make sense to the actual users.

Design lead on a cross-functional squad.

I owned the design end-to-end: information architecture, user flows, high-fidelity prototypes, edge-case mapping, and the component library that engineering followed during the build. I worked closely with the PM on scoping and with the client to validate decisions at each phase.

My focus was turning a fragmented web product into a native app that worked for three user profiles — without building three different interfaces to match.

What each decision looked like in practice.

One architecture, three profiles

Instead of building three parallel apps, I designed a single information architecture where shared entities (shipments, contracts, messages) looked the same for everyone — but surfaced role-specific actions where it mattered. The SuperAdmin kept full control, Carriers and Generators saw only what they needed to act on.

One label = one concept

The old product used different terminology depending on who was logged in. I worked with the PM to collapse these into a single vocabulary. Example: "With Offers" means the same thing for every profile — Offer + Accept/Reject + Counter-Offer. Same term, same actions, same validations.

The happy path wasn't enough

I mapped critical edge cases: rejected documents (view missing requirements, retry, or escalate), incomplete data (validate and ticket), on-route incidents (pause, reassign, or escalate to Operations), and failed closures. Good UX covers both the expected and the unexpected.

Key insights I carry forward.

• Architecture first avoids loose screens and circular discussions. Without a map, every conversation starts from zero.
• One term = one action. Aligning language across profiles speeds everything up — and prevents bugs caused by misaligned expectations.
• Prototype by flow beats waiting for "the complete app." Small validations compound faster than big releases.
• Actively guide the client. The visual map eliminates ambiguity and shortens review cycles.
• What we couldn't validate: early task testing with real carriers and generators. In hindsight, a five-second test on the first action per profile would have saved weeks.