Healthcare software that holds up when the rules get complicated

Mostly the decision about where the model sits. AI works well on the operational layer: processing documents, structuring records, cutting the administrative load on clinical staff. Our Dr. Max leaflet system is a good example, an AI-accelerated build that put a working product in the client's hands in days. AI works badly as an unsupervised decision-maker where a wrong output has a clinical consequence. The projects that succeed start from a specific, measurable operational problem and keep a human in the loop on anything clinical. The projects that fail start from "we should use AI" and go looking for a use case afterwards. If a vendor is enthusiastic about putting a model in a position where nobody checks it, that tells you something about the vendor.

Start with what the staff actually do all day, not with what the software could do. Most clinic digitalisation projects fail because they optimise a process nobody had a problem with while leaving the real bottleneck untouched. The sequence that works: map the current workflow including the paper and the spreadsheets, find the two or three points where time is genuinely lost, and solve those before touching anything else. Then budget for the transition period when the old and new systems run in parallel. That phase is always longer than anyone plans for.

Yes. For Calma we integrated with the medical records system a Slovak clinic is legally required to keep. That system communicates in an unusual way and offers none of the standard connections, so we built it to sync in the background while the patient's booking goes through immediately. For ZITA we integrated with TISS, the ERP system used by ophthalmologic clinics, for real-time stock sync.

Yes, and we're ISO 27001 certified for information security and ISO 9001 for quality management. All client infrastructure runs on servers in Germany, so no data leaves the EU. Every healthcare product we build handles personal data under GDPR as a design constraint from the first architecture decision, not as a review at the end. For Danone we designed the entire data infrastructure around privacy compliance from day one, consolidating scattered sources into a single compliant warehouse.

Central Europe has a concentrated cluster of studios with genuine healthcare experience, working for clients across the EU. The practical advantages are EU data residency by default, GDPR as a native requirement rather than an imported one, and overlapping time zones with Western European clients. BRACKETS is a 20-person product studio based in Bratislava. Our healthcare clients include Dr. Max, Danone, the European Association of Nuclear Medicine and the Vienna-based health venture of UNIQA Insurance Group.

It depends on the scope of work more than on anything else: how many domain rules the system has to hold, how many existing systems it has to talk to, and how much of the compliance work is already done on your side. Two clinics of similar size can land a long way apart. As a range, our healthcare projects typically start around €25,000 and most sit between €50,000 and €200,000. A discovery phase turns that range into a real number and a scope you can hold us to.

Yes, for a defined scope. That budget realistically covers a discovery phase, a working prototype, or a focused tool that solves one clear problem. It does not cover a full patient-facing platform with multiple integrations and compliance work. We would rather scope something small that ships and works than start something large that runs out of money halfway through.

The same thing drives the timeline as drives the price: how many rules the system has to hold and how many existing systems it has to connect to. A discovery phase that ends with a validated scope and an architecture usually takes a few weeks. A first working version that real patients or staff can use is typically a matter of months rather than weeks. After that, healthcare products do not stop. They evolve for as long as the organisation does. Our Danone partnership has run since 2010 and EANM is on its second project with us.

Start by looking at what you can buy. We do. For Calma we evaluated several ready-made booking tools and built a prototype before concluding they couldn't handle session sequencing, credit logic and bulk rescheduling. For EANM, a generic learning platform could not carry specialist certification paths. The right question isn't build versus buy, it's at what point customising a bought tool costs more than owning the thing outright.

You do. That includes the repository, the documentation and the infrastructure. We build so that another team could take over, because a healthcare product will still be running long after the people who wrote it have moved on.

Yes, and often that's the better setup. With Mavie Next we joined as an external development partner supporting a competent in-house team rather than replacing it.

Most of our healthcare work is long-term. Danone since 2010, ongoing SLA support for ZITA covering everything from CMS updates to helping a customer who cannot log in, and a second project for EANM after the first. We treat launch as the point where the product starts, not where the engagement ends.