Answers grounded in a person’s own record
An AI assistant that explains lab results, history and notes in plain language, using the person’s own data and a curated knowledge base. Built with retrieval-augmented generation and running in the Healsens app.
Personal, explainable, within clear limits
The assistant knows the record it is talking about and stays within what it is allowed to do.
Grounded in the record
Answers use the person’s own results and history, not an average from the internet.
Changes over time
Because the record is longitudinal, the assistant can explain trends, not only single values.
Agent routing
A routing agent picks the right specialist step; tools fetch tests and records when they are needed.
Curated knowledge base
Background comes from a vector knowledge base built from approved content, which can be your own.
Clear boundaries
It explains and organises. It does not diagnose or give treatment advice, and points to a doctor for medical questions.
Model choice
The model is chosen per use case and can be changed. What it may see is defined by the tools, not the prompt.
Retrieval-augmented generation on personal data
Ask
A question in plain words, such as “How has my vitamin D changed since my last test?”
Retrieve
The assistant fetches the relevant results from the record and passages from the knowledge base.
Answer
A language model writes the answer from that material, with the person’s own values.
Stay in scope
Medical questions are pointed to a doctor. The assistant does not make clinical decisions.
An assistant on your data and your knowledge base
The assistant runs in the Healsens app today. We build the same architecture into your product or project: your records, your approved content, your hosting.
- Your data modelTools are built around the records you hold.
- Your contentA knowledge base from material you approve.
- Your environmentHosted in your cloud or on EU infrastructure we run.
Where it fits
Explaining results
Help users understand their own lab results and how they change over time.
Participant support
Answer participants’ questions about their own study data, within the study’s rules.
Assistants in your product
Add a grounded assistant to your product without building the retrieval and agent layer yourself.
Frequently asked questions
Which models do you use?
We choose the model per use case and can switch. What matters more is what the model is allowed to see and answer.
Can it use our own knowledge base?
Yes. The knowledge base is a separate, curated collection that we build from your approved content.
Does it give medical advice?
No. It explains and organises a person’s own data and suggests discussing medical questions with a doctor. Healsens is not a medical device.
Let’s talk about your data
Tell us what you are working on. We reply by email and, where it fits, suggest a short call to look at a sample together.