Services & commerce
Healthcare
Software for the administrative weight around care — scheduling, records, and correspondence — with a conservative line on anything clinical.
Context
What this sector actually looks like
The clearest wins in healthcare are administrative. Scheduling, reminders, intake, correspondence, and records retrieval consume clinical time without requiring clinical judgement, and they are well suited to automation.
Anything touching diagnosis or treatment is a different category of risk, requiring regulatory approval we are not positioned to provide. We build systems that support clinicians and we do not build systems that make clinical decisions.
Patient data raises real constraints on where processing may happen. That usually points towards on-premise or region-locked deployment, which is a design decision to make at the start.
The hard parts
Where it usually breaks
Problems worth naming before proposing anything to fix them.
Administrative load
Clinical staff spending significant time on scheduling and paperwork.
Record retrieval
Patient history spread across systems and paper.
Appointment management
No-shows and manual rescheduling absorbing front-desk capacity.
Data residency
Constraints on where patient data may be processed, often ruling out hosted APIs.
What we build for it
Where software earns its cost here
Scheduling and reminders
Booking, confirmation, and rescheduling handled automatically over the channels patients use.
Intake digitisation
Forms captured as structured data rather than scanned paper.
Records search
Retrieval across historical records, with the source document always shown.
On-premise deployment
Open-weight models running inside your infrastructure where data cannot leave.
FAQ
Questions people actually ask
Do you build diagnostic tools?
No. Clinical decision support is a regulated category requiring approvals we do not hold, and building it without them would be irresponsible. We work on the administrative layer around care.
Can patient data stay on our servers?
Yes. Where residency requirements rule out hosted model APIs, open-weight models run on your infrastructure. We will be explicit about the quality trade before you commit.
Can you integrate with our existing system?
Where it exposes an interface, yes. Many clinical systems are closed, and we assess that honestly before proposing anything.
Start here
Tell us what is slow, manual, or breaking.
Answer a few questions and get a written brief back — scope, proposed architecture, and what it would take to build.