Versions
More time for patients on every ward
Doc Dialog Clinic brings AI-powered documentation to the hospital, from admission history to discharge letters. Our AI suite is also available offline and supports the creation of discharge letters, transfer reports, and other clinical documents. Interdisciplinary, secure, with data in Switzerland.
- Offline-capable AI for discharge & transfer letters in minutes
- Interdisciplinary: doctors, nursing & therapy in one tool
- Multilingual, error-free reports at Swiss level
- Connects to existing hospital systems, data in Switzerland
Members & partners
Clinical everyday
Relief across every profession
Discharge & transfer letters
From notes and progress data, structured discharge and transfer letters are produced in minutes instead of hours.
Interdisciplinary & multilingual
Doctors, nursing and therapy document in the same system. Even foreign-language staff phrase things flawlessly at Swiss level.
Secure integration
Connection to existing hospital information systems. Data stays in Switzerland and is fully DSG-compliant.
Modules
Modules for the clinic
Documentation
Discharge letter
Structured, professional discharge letters from progress notes and keywords, ready for sign-off.
Communication
Consult & transfer reports
Fast, precise reports for referrers and follow-up institutions, transferred with one click.
Ward
History & progress
Dictation or keywords become clean entries, ready for your hospital information system.
How it works
From dictation to clinical report
Capture
Dictate or jot keywords right on the ward.
AI structures it
Doc Dialog produces professional discharge letters, transfer reports, and other clinical documents, including offline when required.
Review
The care team reviews, adds and signs off.
Transfer
Straight into the hospital information system or to Word/PDF.
🇨🇭 Patient data remains in your practice system. Doc Dialog does not store patient information long term; data is processed only in transit and then handed over to your PIS.
The big wins
The strongest statements from the pilot
Case 1 carries the pilot
PDF import and summarization is the clear headline case: roughly 30 minutes become about 5 minutes per case. The process is already running productively in the secretariat, with status fully met / in operation. According to the pilot assessment, this value case alone already exceeds the cost of the entire POC.
The architecture is proven
The on-prem local server with open-source models runs inside the Zurzach network without an internet connection and without token costs. At the same time, the pilot ran technically stable in parallel operation across all tested cases.
The two-environment model works
Patient data stays local, while the online test environment is used for fast iteration. This made it possible to implement 28 new requests during the pilot and then pull them down locally. This exact model kept development fast without getting stuck in data-protection disputes.
The pilot generated more scope than it consumed
The workshops produced 13 use-case clusters, 3 were contractually prioritized for the POC, and 4 genuinely new cases emerged in real use: REI, doctor search, PubMed/Neuro PDF, and rehab trajectories.
Secretariat
This is the strongest evidence in the whole pilot
Time & throughput
In the secretariat, around 25 minutes were saved per case, at roughly 30 cases per day. That means more document throughput per staff member and measurable day-to-day relief.
Holiday cover became feasible
"Without the tool, we would not have managed holiday cover." This feedback is stronger than a pure time measurement because it describes a real operational outcome felt directly in daily work.
Task shifting from physicians to secretariat
"Meaningful shifting from physicians to the secretariat." Work that previously consumed physician time now stays in the secretariat. This makes physician time relief a second, often unbudgeted saving on top of secretarial time.
The unglamorous but important OCR win
Protected, non-copyable, and scanned PDFs suddenly became processable. That removes an entire class of daily friction: fewer follow-up requests, fewer "please send it again as text" emails, and less manual searching.
Transferable mechanism
Why this works and what other hospitals can take from it
The lever is pre-work, not just text
The biggest lever is not "AI writes text" but standardizing and accelerating the administrative pre-work around documents.
Maturity follows input structure
Structured, repeatable inputs such as referral PDFs become productive quickly. Free speech in a room needs more setup. Same models, same app, but completely different operational maturity.
The dictation blocker is recording quality
Distance, volume, background noise, speaker changes, and missing golden-truth recordings are currently the blockers, not primarily the AI models. This is a hardware and process problem that is often mistaken for an AI problem.
Other hospitals should start in the secretariat
That is where the pain-to-risk ratio is best: high value, low clinical liability, and measurable results within weeks.
Cloud for iteration, local for data
Anyone forcing development and patient data into one environment usually gets either slow development or a data-protection problem. The split model was a central success factor.
Interfaces are not the first bottleneck
Second-screen use and copy & paste into the HIS were rated sufficient in the pilot. Interfaces can come later. That removes the usual 12-month IT blocker from the critical path.
Honest caveat
The credible claim is deliberately narrow and strong
Case 1 is strong and already robust today: document-heavy secretarial work on an on-prem basis pays back within the pilot. Case 2 shows good text quality, but the operational test is still outstanding. Case 3 shows potential, but remains open for everyday clinical use. This honest boundary is exactly what protects the credibility of the strong PDF/document value case.
Become a pilot clinic
Doc Dialog Clinic is being built together with hospitals. Help shape it and secure early access.