WHAT WE DO

FORTIAID

Clinical answers that do not need a network or a wall socket.

A health worker in a district clinic has a patient in front of them and a question they cannot answer from memory. The guidance that would answer it sits on a server they cannot reach, because the line is down, the power is out, or neither ever arrived.

FORTIAID is a sealed case that carries the guidance and the reasoning with it. It runs on its own battery, recharges from a folding solar panel, and answers questions in ordinary language without touching a network at any point. Nothing leaves the case.

THE PROBLEM

Where the infrastructure ends, so does the evidence

Conflict, storms, attacks on hospital systems, and simply being a long way from a city all produce the same situation: clinical staff working without the references their training assumes will be there. The gap is not clinical knowledge. It is access to that knowledge at the moment a decision is being made.

Networks fail first
When infrastructure is damaged, overloaded or deliberately cut, clinical systems go down with it. The tools that depend on a connection stop being tools exactly when the caseload rises.
Paper cannot keep up
Printed protocols are already behind by the time they are distributed, revisions rarely reach the last clinic in the chain, and a full set of them is heavy to carry to a site that has to be reached on foot.
Power is the binding constraint
Equipment that needs a mains socket is no use where the absence of mains power is the problem. Anything meant for the field has to carry its own supply and earn every watt it draws.

THE DEVICE

What is in the case

A hand-carried case, sealed to IP67 against water and dust, at roughly six kilos including everything below.
A folding solar panel and a battery bank, so the unit runs a full clinical day away from a socket and recharges in daylight.
A compact computer that does all of the reasoning on the device itself. No connection is required at any point in normal use.
A library of treatment guidance from WHO, ECDC and national health authorities, held locally and updated deliberately rather than silently.
A screen and voice input, so the device can be used by staff who would rather ask a question than type one.

HOW IT WORKS

An answer you can check, not an answer you have to trust

A clinician acting on an answer needs to know where it came from. FORTIAID is built so that the source is always available, and so that the device declines rather than guesses.

Grounded in published guidance
Answers are drawn from the guideline library held on the device, not from whatever a language model happens to have absorbed in training. Where the library does not cover a question, the device says so instead of producing something plausible.
Every answer shows its source
The passage an answer was built from is shown alongside it, so a clinician can read the underlying guidance and judge it, rather than take a summary on faith.
A safety filter that escalates
A defined set of critical presentations is checked on every query. Those are routed to human judgement rather than answered, because the cost of a confident wrong answer in that category is not recoverable.
Built for the room it is used in
Local languages, and voice input, so that reading and typing are not what stands between a health worker and the answer.
Nothing leaves the device
No patient information is transmitted, because in normal use there is nothing to transmit it to. Data privacy here is a property of the architecture rather than a policy written on top of it.

(A) Data can be collected from various sources, such as the internet, or (B) specific document pools.
(C) This collected data is processed and prepared for analysis. An embedding model is used to convert the data into vectorized representations (numeric form), making it easier for AI to understand and process.
(D) The vectorized data is stored in a vector database, a special type of database optimized for searching and retrieving this kind of information efficiently.
(E) Over time, the large language model can be fine-tuned using the data from the vector database to improve its accuracy and relevance in generating responses.

(1) A user inputs a query, and the system first analyzes and understands the question.
(2) The system then converts the query into an embedding (a vectorized form), similar to the data stored in the vector database.
(3) The query embedding is used to search the vector database and retrieve relevant information based on similarity to the user’s question.
(4) The retrieved data is combined with a large language model (LLM) to provide deeper insights and context (5).

FIELD VALIDATION

Thirty days a site, in the settings it was built for

Units were placed with clinical teams in low-resource and conflict-affected settings and run for thirty days at each site, on solar and battery power alone. Every query was logged. A blinded sample of those logs was reviewed by the Research Unit for General Practice at the University of Southern Denmark, which had no role in building the device.

Measured in the field pilot	Result
Clinical precision, on a blinded sample reviewed by SDU	91 %
Uptime across the pilot period	over 98 %
Clinical queries answered	more than 3,400
Clinicians trained on site	more than 60
User satisfaction after thirty days of daily use	4.4 out of 5

Figures from the 2026 field pilot. Independent clinical review by the Research Unit for General Practice, University of Southern Denmark. Results describe performance in that pilot and are not a guarantee of performance in a different setting.

See it answer something

The quickest way to understand FORTIAID is to ask it a question you already know the answer to, and then one you do not. We can bring a unit to you, or walk your clinical team through the field data and the way the validation was done.

Book a demo