CIS stability meets AI speed: why Swiss hospitals should not wait to adopt AI productivity tools
Updated: 3 days ago
The key question about AI use in hospitals is no longer whether it will come, but where it will first be made available: in the release cycle of the clinical information system (CIS), or in a specialized AI productivity layer that can already be integrated in parallel today?
For hospital management, this is not a technology question but a question of economic prioritization. Anyone who waits for the core system to deliver administrative relief is postponing not only innovation, but also capacity, process stability and measurable economic benefits.
Economic pressure and administrative burden do not allow any delay
The PwC hospital study 2025 shows that the margins of acute care hospitals have improved slightly but remain below the level needed in the long term.[1] At the same time, the Swiss healthcare system cost around CHF 94 billion in 2023.[2]
At the same time, the administrative burden remains a structural problem: the Federal Office of Public Health (FOPH) notes that administrative effort in the healthcare system has been heavily criticized for years and that targeted relief measures need to be developed.[3] The FMH shows how high this burden is: hospital doctors in acute somatic care spend an average of 114 minutes per day on documentation.[4]
Anyone deciding on AI productivity is therefore not talking about a nice-to-have, but about whether highly qualified professionals spend their time on administration or on patient care.
Switzerland: strong in care delivery, demanding in coordination
Switzerland has a dense and high-performing care network. This is a strength and at the same time a structural challenge. A significant share of administrative friction does not arise within individual specialist processes, but at the transitions between institutions and roles.
Info box: the Swiss care landscape (selection[5-10])
Area | Number |
Hospitals | 275 |
Medical practices and outpatient centers | 17,631 (2022) |
Home care (Spitex) providers | 2,971 (2023) |
Retirement and nursing homes | 1,474 |
Pharmacies | 1,830 |
Psychological psychotherapy practices | 4,788 (2023) |
Physiotherapists | approx. 18,000 (2020) |
Mandatory health insurance (OKP) insurers | 37 (+5 daily allowance insurers) |
A Swiss study among general practitioners in central Switzerland shows: 88% expect the discharge report within 48 hours, but only around 30% report that they are actually informed within this time.[11] The result: follow-up questions, additional work, duplication and uncertainty in aftercare.
The clinical information system keeps operations running, specialized AI accelerates progress
Clinical information systems are generalists. As the system of record, they cover patient administration, service recording, nursing, laboratory, billing and dozens of other processes. This breadth is their strength, and at the same time their structural limit when it comes to the speed of innovation.
The CIS is the reliable long-distance runner: enduring, stable, proven. The sprint toward AI-powered productivity requires a specialist that is focused on administrative relief and optimized for speed and iteration.
CIS vendors must simultaneously deal with regulation, legacy code, interoperability and the requirements of hundreds of customers. The result is semiannual or annual release cycles. A specialized AI solution can release updates monthly, translating progress in the underlying AI models and their targeted orchestration directly into everyday clinical practice.
The evidence: the benefits of AI are measurable today
Clinical evidence: USZ simulation study
A simulation study by anesthesiologists at the University Hospital Zurich, published in the Journal of Clinical Anesthesia, examined the use of Isaac’s AI assistant in preoperative consultations (Rahrisch A et al., 2026).[12]
The results:
18% shorter total consultation time, due to faster documentation, not shorter conversations
78% less screen time: doctors looked at patients, not at screens
87% fewer keystrokes
21% better doctor-patient interaction (self-assessment)
16% lower mental workload
60% of doctors preferred AI support to secretarial assistance
(Selection; you can find details in our blog article )
Practice data from PoCs and routine use [13]
Across pilot projects and routine use, consistent results emerge in four dimensions:
Capacity & productivity | 51–74% time savings per report 1–2 hours gained per working day 5–10 hours of capacity gained per healthcare professional each week |
Staff retention | 75–88% of users experience reduced stress |
Finance & OPEX | Measurable benefit in CHF from time savings and additional capacity Reduction in overtime and transcription costs |
Digitalization & processes | NPS consistently above 8.5 High usage and retention rates Clear preference for AI-supported over conventional ways of working |
What does this mean financially?
Sample calculation: a hospital with 60 residents and senior physicians, with average time savings of 7 hours per week per doctor:
60 doctors × 7 hours/week × 46 weeks = 19,320 hours per year
This corresponds to around 9.2 full-time equivalents (based on 2,100 annual hours).
This capacity is available for additional patient care, reducing overtime, research and teaching, or strategic projects.
We would be happy to work with you on concrete ROI calculations based on your hospital’s key figures.
What you are right to question
“Another tool next to the CIS: isn’t that exactly the fragmentation we want to avoid?”
A legitimate concern. The difference: Isaac is not an isolated solution meant to sit permanently alongside the CIS, but a bridge. The solution delivers value as a standalone tool today and is being integrated step by step into CIS systems. The goal is not permanent coexistence, but convergence. The disruption for doctors remains minimal: Isaac runs on smartphones and computers, and the results can be transferred directly into the CIS.
“How do you make sure the tool is actually used?”
Adoption is the decisive success factor, and the most common reason digitalization projects fail. The data speak for themselves: NPS above 8.5, high retention, and in the USZ study 60% of doctors preferred AI support to secretarial assistance. Isaac solves a problem that doctors feel every day. Saipient supports the rollout with structured on-site onboarding, active support in daily work and transparent usage measurement.
“What happens if the AI makes mistakes? Who is responsible?”
AI-generated documentation is a draft, not a final product. Every report is reviewed and approved by the doctor before it is transferred to the CIS. Clinical responsibility always remains with the treating professionals. Isaac supports; the tool does not decide.
“What about data protection?”
Saipient is a Swiss solution. Patient data is processed and stored exclusively in Switzerland. The solution complies with the revised Swiss Federal Act on Data Protection (revFADP). Patient data is not used to train AI models and is deleted after 96 hours. On request, a sovereign cloud solution is also available together with our partner PHOENIQS. For a more detailed discussion, please see our blog article “Digital sovereignty in healthcare”.
What really matters to the C-suite
In discussions with hospital management, recurring requirements emerge:
Data protection: non-negotiable, especially for Swiss health data
CIS integration: not as a prerequisite, but as a clear development path
Process optimization: digital and embedded in the work process, not technology for its own sake
Platform thinking: not another isolated tool, but a solution with a convergence perspective
Measurable KPIs that interest the C-suite:
CMI-adjusted length of stay
Material cost intensity
Personnel cost intensity
Integration into existing work processes
User-friendliness and acceptance
These requirements are justified, and they do not argue against starting immediately. They argue for a provider that can measure these KPIs transparently, guarantees data protection in line with Swiss standards and brings a clear integration roadmap.
Recommendation: start now, integrate in parallel
For CEOs and CFOs of Swiss hospitals, the decision logic can be condensed into three points:
Not “whether AI” but “how fast” determines competitive advantage. Administrative pressure is rising, the shortage of skilled staff is worsening and the documentation burden is growing. Every month without AI support is a month of lost capacity.
Pilot projects are the low-risk entry point. Measurable results in weeks, not years. No complex IT installation, no dependence on the CIS, no a multimillion-franc investment
The CIS roadmap is not a reason to wait, but a reason to start in parallel now. AI-first today, integration tomorrow, convergence in the medium term. Not a contradiction to the CIS strategy, but an acceleration.
Conclusion
The CIS remains the backbone of the hospital: the reliable long-distance runner that supports clinical operations. The sprint toward AI-powered productivity, however, requires a specialist that already delivers today.
The evidence is there: less documentation time, less screen time, less stress, more capacity, better doctor-patient interaction. The legitimate concerns can be addressed: data protection in line with Swiss standards, a clear integration path, transparent quality development and measurable adoption.
The question for the C-suite is not: “Is the technology ready?” The technology is ready. The question is: “Are we ready to realize the benefits now?”
Saipient supports Swiss hospitals in evaluating and implementing AI productivity solutions, regardless of the CIS in use.
Footnotes / references
[1] PwC Schweiz. Spitalstudie 2025: Schweizer Spitäler – So gesund waren die Finanzen 2024. 2025.
[2] Bundesamt für Statistik (BFS). Kosten und Finanzierung des Gesundheitswesens im Jahr 2023. 2025.
[3] Bundesamt für Gesundheit (BAG). Administrativer Aufwand im Gesundheitswesen. 2025; Kurzbericht administrativer Aufwand im Gesundheitswesen, 2024.
[4] FMH / Schweizerische Ärztezeitung. Administrative Entlastung erforderlich. 2024; FMH-Befragung zur Arbeitsrealität in der Akutsomatik.
[5] Gesundheit Taschenstatistik 2025, BfS
[11] Tomaschek R, et al. A survey among primary care practitioners in central Switzerland on hospital discharge communication. Swiss Medical Weekly. 2025.
[12] Rahrisch A, Braun J, Ganter MT, Roche TR, Spahn DR, Tscholl DW, Kaserer A. Generative AI in preanesthetic consultations. Journal of Clinical Anesthesia. 2026.
[13] Saipient. Internal PoC data and ROI analysis for Isaac; USZ/Saipient analysis, on file.
