From AI model to reliable healthcare application with MLOps at LUMC
LUMC
Care.ai
Digital healthcare assistant for reimbursement information.
Give employees faster access to relevant information about reimbursements.
Trusted by leading organisations
Why we built this,
Staff at pharmacies and medical device suppliers look up reimbursement information every day. That information is spread across different sources, varies by health insurer and changes regularly.
Answering a question therefore takes considerable research and specialist knowledge. Staff consult multiple documents and often rely on experienced colleagues for complex questions. This takes time and can mean patients do not always get the same answer.
Care.ai makes reimbursement information accessible through a single application. Staff ask questions in everyday language and receive answers based on the connected sources.
What you can expect,
A single point of access for reimbursement queries
Care.ai makes information from various health insurers about pharmaceutical products and medical devices accessible. This means staff no longer have to search each source separately.
The information the assistant can use depends on which sources are available and connected.
Answers based on your own knowledge sources
The assistant searches the selected documentation for information relevant to the question. Where possible, source references show what an answer is based on.
You decide:
- Which sources the assistant may use
- What types of questions the assistant is used for
- Who has access to the application
- When a question needs to be reviewed by a member of staff
Support for staff
Care.ai does not replace professional judgement. The application helps staff find relevant information more quickly and answer questions more consistently.
In unclear, exceptional or medically complex situations, a qualified member of staff still needs to review the answer.
Our approach,
-
Start with a clearly defined issue
We first identify which reimbursement questions take a lot of time and which sources are needed to answer them. This gives us a clear use case that fits into employees’ day-to-day work.
-
Connect sources and test answers
We make the selected information accessible to the assistant and test its answers with employees who know the subject matter. We check not only whether an answer sounds plausible, but also whether it is consistent with the available source information.
We use questions the assistant cannot answer adequately to refine the use case further. -
Introduce it in a controlled way
We agree on access, use and human oversight. We also establish who is responsible for keeping the connected information up to date.
After the initial launch, the assistant can be expanded with additional sources or new types of questions.
Not a general-purpose chatbot, but a focused assistant that helps staff find and apply reimbursement information.
The assistant helps staff answer questions, while they remain responsible for the final assessment.
What we've built before,
Three projects from our work with data and AI.
Further reading
Blog posts and in-depth articles from our team.
24 Aug 2026
Three lessons from our AI projects
What we learned from dozens of AI projects: start with the decision, treat data quality as an organisational issue and arrange management in advance.
18 Aug 2026
From pilot to production: why most AI projects stall
Most AI projects stall not because of the technology, but because of data, teams and processes. Read what it takes to move from pilot to production.
25 Jun 2026
Responsible AI: from principles to practice
AI is developing rapidly. New models emerge almost every week, and more and more organisations are experimenting with AI.
Get in touch
Stef Stoelwinder
Business DevelopmentE-mail info@datasciencelab.nl
Phone +31 (0)20 3636 724
Frequently asked questions,
The assistant focuses on questions about health insurers’ reimbursement policies for pharmaceutical products and medical devices. Its exact scope depends on which sources are connected and the agreements made about its use.
The assistant uses information from the connected sources. It therefore needs to be clear who manages these sources and how changes are processed. The solution can only provide up-to-date answers if the underlying information is current and accessible.
The application helps staff find and interpret information. Exceptions, unclear questions or situations with more significant consequences still require assessment by an authorised member of staff. We determine the appropriate checks based on the question and the risks involved.
A narrowly focused assistant uses selected sources and works within a clearly defined subject area. This makes it easier to check which information is used and when a question falls outside the application’s scope. It also makes the answers easier to manage, but human oversight remains important.
Reliable source information, clear responsibilities and agreements on human oversight provide the foundation. You also need to define who has access, how answers are checked and what happens when the assistant is not sufficiently certain. This makes responsible use part of the application itself.