AjudAí MVP
Hospitals discard medication that's about to expire while others, in the same city, run short of stock. I designed the AjudAí MVP in 6 months — a platform that connects hospital pharmacies to exchange and donate medication.
Context
AjudAí started from research at the University of Sorocaba. Tracking medication management in hospitals, the researchers saw the same pattern repeat: consumption fluctuates and stock never keeps up — sometimes there's surplus that expires, sometimes there's shortage and the patient's treatment stops. That observation gave rise to the startup, to get hospitals in the same region to cooperate and manage stock together through exchange, loan and donation of medication.
Problem
AjudAí grew out of a logistics problem. Hospitals discard expired medication while others, in the same region, run out of that same item. Two stocks that never balance out, and the loss is costly: in public hospitals, expiration consumes about 2% of stock every year.
Being an MVP, the platform's first job wasn't usability — it was market: proving that waste of this size could be tackled, with real value for hospitals. Before any interface question, there was a product problem to solve.
Process
The start
Architecture
Interface
Iteration
Solution
Two profiles with different needs, a 6-month deadline and a small team called for decisions on where to focus effort — not a single product trying to serve everyone.
Two profiles, two products
AjudAí had two opposite profiles: the manager, who registers surplus and looks for what's missing, and the administrator, who keeps the platform running. Distinct goals called for distinct products, each with its own interface and vocabulary. The exchange-between-strangers mechanic came from OLX and Enjoei — with medication in place of used goods — and it defined how to communicate an offer and build trust without a physical intermediary.
From those profiles, I structured the information architecture. The sitemap splits the home into two lists — what's missing and what's surplus — and gathers management under "my medications." The registration flow starts with a choice between offering or requesting, moves to the form, and blocks incomplete entries.
UI design of the registration flow
From the medication list to registering an available or needed item, in four steps.
Design system as a product decision
A 6-month deadline, a small development team. Choosing an established design system wasn't about saving effort — it was about deciding where to concentrate it.
Ant Design's ready-made base covered what was commodity and freed up time for me to design what was specific — flow and architecture, where the risk of error was greatest. It also closed the gap between design and product: documented, tested components that the team implemented directly, with no rework.
Responsiveness is access, not aesthetics
Hospital managers don't always work on desktop. Making sure the platform worked across different devices wasn't an aesthetic concern — it was an access decision for the real user profile, who opens the system from wherever they are, on whatever device is at hand.
User testing
Usability tests were remote: I sent the Figma prototype link and ran each session over Google Meet. The user received the task to carry out and navigated the prototype while I noted what could be improved. Those sessions produced the final adjustments that closed the MVP.
Results
MVP delivered in 6 months, within scope and ready for the first users and investors. I left before the usage measurement cycle, so instead of metrics I never tracked, I record what could already be stated and how each goal would be measured.
-
Whether waste was reduced
Drop in expiration-related disposal and volume of medication reused through exchange or donation, compared against stock history before and after.
-
Whether communication expanded
Number of active institutions and of exchanges completed between them, along with the response rate to requests.
-
Whether consumption got simpler
How far ahead the shortage-or-surplus alert fires, and the effort to register and locate an item, measured in time and steps in the task.
Credibility
Pharmacists responsible for stock management described the platform as intuitive, simple and practical — exactly what you'd expect from a product for healthcare professionals without a technical background. Usefulness between institutions in the same region confirmed local cooperation as the product's central hypothesis.
Credibility also came from institutional recognition: FAPESP, Sebrae for Startups and UFRJ feature as partners, alongside UNISO, where the project originated. FAPESP's support signals scientific merit evaluation; the partnership with Sebrae, that the product was treated as a business — not just as research.
Learnings
Scope protects delivery
Several proposals came up to expand the platform — native chat, contract issuance, an alert system. All made sense for a mature product. I prioritized the functional core within the deadline and documented the rest as roadmap. A useful MVP is worth more than an incomplete, ambitious product.
Design explains itself, not just delivers
The team included researchers and developers with no prior experience with digital product process. Making each design step visible and understandable — not as bureaucracy, but as alignment — was what prevented rework at the end of the cycle.
Choosing the design system is a product decision
Choosing Ant Design wasn't just a visual choice — it was a delivery decision. Designing with components the team could implement directly eliminated the gap between what I delivered and what reached the final product, with no visual debt or rework between design and dev.