From 6 loose applications to a single database.
FundaCáncer is a Costa Rican organization that supports cancer patients from Hospital Dr. Rafael Á. Calderón Guardia. Before, each volunteer logged cases in separate applications with no connection between them — data traveled from one computer to another by hand, in exported files. Today everything lives in a single real database, in production.
Data typed in by hand is almost never complete.
Duplicates nobody notices
The same person logged twice, with the name spelled differently each time — until someone tries to cross-reference the data and it doesn't line up.
Empty fields that should be required
A date, a phone number, a status left blank because whoever entered it didn't have the data on hand at that moment — and nobody ever goes back to fill it in.
Inconsistencies between different sheets or files
The same case described differently in two places, because each person logged it on their own, without seeing what the other had already entered.
Nobody finds out until it's too late
With nobody reviewing the database every day, a data-entry error piles up for weeks before anyone notices it — usually once it's already caused a problem.
The same idea, applied to the entire database.
It's not a pretty dashboard nobody checks — it's an AI layer that watches the data and acts before the problem grows.
Day Panel
Every morning, the system decides on its own who to attend to first: today's appointments, recently discharged cases with no follow-up, and overdue cases — without anyone flipping through notebooks or spreadsheets.
Compliance semaphore
Watches each case's official deadlines and flags what's about to expire in red, yellow, or green, so someone can act before it becomes a problem.
AI that audits the database itself
The same assistant that chats with the team can review the data in the background: duplicate names, empty fields, inconsistencies between records — and flag them before anyone else notices.
Reports that no longer get built by hand
The metrics that used to get assembled once a semester in a slideshow are now available anytime, calculated live from the same database.
These alerts can arrive over WhatsApp or email — the same channel where Mercurio already lives, with no extra app to check.
The right numbers change decisions, not just decorate a report.
A KPI that updates itself isn't decoration — it's the difference between guessing where time or money is short, and knowing for sure.
Catches patterns before a person notices them
When the data is complete and current, a trend that would take months to spot by eye — a zone with more delays, a type of case that keeps getting complicated — shows up on its own in the panel.
Focuses effort where it actually pays off
Instead of spreading time and budget evenly across everything, the numbers show where a bit more attention today prevents a bigger — and more expensive — problem later.
Backs every investment decision with data
Asking for more budget, justifying a hire, or defending a project in front of a board is a different conversation when the number comes from a live system, not an estimate.
Automation is no excuse for lowering your guard.
Data isolated by organization
Each organization's data is kept separate at the database engine level, not just the application — a code bug can't leak information from one into another.
Every action gets logged
Who did what, when, and on which case, with an audit trail — not just trust that nobody makes a mistake.
The AI doesn't train on your data
When the assistant needs a language model, the connection requires that no provider retain or use that data to train its own models.
Built with Costa Rican law in mind
Personal data handling follows the principles of Costa Rica's Law on the Protection of Individuals from the Processing of their Personal Data, not just generic best practices.
“We accompany every patient with commitment, warmth, and hope at every step of the way.”
Does your business also depend on loose spreadsheets?
Tell us how you handle your data today and we'll tell you, straight up, whether it's worth automating.