German healthtech runs on GKV procurement and BfArM listings.

Germany operates the most institutionally dense healthcare market in continental Europe and the third-largest medical-device industry globally. The structural baseline is a dual-payer Bismarck system: roughly 74 million residents insured under the statutory Gesetzliche Krankenversicherung (GKV) — about 90% of the country — plus 8.7 million in private Vollversicherung (PKV).

GKV is administered by ~95 Krankenkassen, with the AOK regional family covering ~27M insured, Techniker Krankenkasse ~11M as the single largest fund, Barmer ~8.68M, and the DAK / IKK / BKK families forming the procurement counterparties any vendor selling to SHI-insured patients must eventually face ([Germanpedia 2025](https://germanpedia.com/top-german-public-health-insurers/)).

The provider side is structurally fragmented and financially fragile. Destatis reports 1,874 hospitals at year-end 2023, of which ~1,585 are acute-care, split roughly public 28% / non-profit 39% / private 33%. The Roland Berger Krankenhausstudie 2025 found three out of four hospitals closed 2024 in deficit, with ~89% of public facilities in serious distress ([PMC 2025 review](https://pmc.ncbi.nlm.nih.gov/articles/PMC12629149/)).

The Lauterbach-led Krankenhausreform (2024-2026 staged implementation) consolidates fewer sites into Level-3 tertiary nodes and reroutes simpler interventions to ambulatory care — a procurement disruption with a 24-36 month tail every healthtech vendor must price into its enterprise pipeline.

Ambulatory care is more granular still. Around 270,000 SHI-billing physicians work across ~100,000 Praxen, scheduled through Doctolib (100,000 DE clinicians, 25M patients), Jameda (CGM-owned), and a long tail of practice-management systems.

Three vendor families anchor the digital-health installed base: CompuGroup Medical (Koblenz) for ambulatory PMS and the eRezept processing layer; Doctolib for scheduling, voice (Aaron.ai, acquired May 2024) and clinician collaboration (Siilo, 2025); AGFA HealthCare, Dedalus, Visus, Sectra, x-tention for hospital information systems alongside Cerner and Epic.

The medtech topline is dominated by anchors most German buyers can name without thinking. Siemens Healthineers (Erlangen) reported EUR 22.4B revenue in fiscal-year 2024, with EUR 12.3B from Imaging alone, ~72,000 employees, and an estimated 30% global imaging share. Carl Zeiss Meditec (Jena, ophthalmology + surgical), B.

Braun (Melsungen, infusion + dialysis), Drägerwerk (Lübeck, anaesthesia + ventilation), Fresenius Medical Care (Bad Homburg, dialysis), and BioNTech (Mainz, mRNA + AI oncology via InstaDeep) round out the backbone, with a thicker Mittelstand tier (Karl Storz, Brainlab, Otto Bock) supplying everyone.

The DiGA framework is the structural feature with no equivalent in any other major EU healthcare market. A BfArM DiGA-Verzeichnis listing under SGB V §139e turns an app into a prescribable, GKV-reimbursable digital therapy.

The directory listed 44 permanent + 14 provisional = 58 active DiGAs as of June 2025, with ~73 cumulative listings since the 2019 Digital-Versorgung-Gesetz introduced the fast-track ([BfArM DiGA-Verzeichnis](https://diga.bfarm.de/de); [Sifted DiGA reality check](https://sifted.eu/articles/diga-promised-german-healthtechs-access-to-73m-patients-but-insurer-roadblocks-and-slow-adoption-are-limiting-its-potential)).

Areza does not build SaMD, does not train diagnostic AI, does not own clinical decision-making, and does not wrap clinical decision support. Anything that would be Software-as-a-Medical-Device under EU MDR Class IIa, IIb, or III, or a high-risk AI system under the AI Act Annex III, is out of scope.

That carve-out is structural, not cosmetic — it keeps Areza out of MDR classification, out of AI Act Annex III high-risk obligations, and out of HWG §3 misleading-claims liability. We partner with Aidoc, mediaire, Smart Reporting, Stratipath, ContextVision, and the clinical-AI cohort. We do not replace them.

Foundation — bilingual DE-EN healthtech / medtech site engineered to signal clinical-validation receipts (peer-reviewed citations, BfArM listing IDs, MDR certificate numbers, CE-IVD class), DiGA-listing-status badges, GDPR Art. 9 readiness language, hosting jurisdiction disclosed, and AI-search-extractable structured data.

The Sie-register default holds for hospital and Kassenärztliche Vereinigung audiences; Du is acceptable for D2C DiGA marketing where the segment expects it. Service pages per Indikation × Stadt for D2C clinics and physio, MSK, dental, and aesthetic chains.

AI Search — capture citations in ChatGPT, Perplexity, Gemini, and Google AI Overviews for German-language buying queries: 'DiGA [Indikation]', 'digitale Therapie [Erkrankung] Deutschland', '[Erkrankung] App auf Rezept', '[Stadt] Klinik [Verfahren] KI', 'Kassenleistung [App-Name]'. Today these SERPs are dominated by the BfArM list, Krankenkasse PR pages, Apotheken-Umschau, Netdoktor, Onmeda, jameda, and a thin layer of Zava content.

Google AI Overviews rolled out in Germany on 26 March 2025, and Health is one of the highest AI-Overview-share verticals globally at ~43% — the citation surface is rebuilding in real time ([projx.de AI Overviews Germany rollout](https://www.projx.de/en/magazine/ai-overviews-rolled-out-in-germany-massive-changes-expected-for-seo-and-traffic)).

Voice Agent — appointment intake, DiGA onboarding, and pre-visit symptom collection in a consent-aware multi-language DE-EN-TR-RU-AR configuration. Turkish and Russian are non-trivial in urban catchments; Arabic matters in the largest cities and refugee-care contexts.

Strictly factual routing (which clinic, when, what to bring, which document) and never differential-diagnosis claims. Hard guardrail: the agent must not give clinical advice or triage to acute care; it routes to the practitioner and logs consent before any voice recording. Doctolib's Aaron.ai acquisition (May 2024) is the precedent that this category is now reimbursable-adjacent in Germany.

Workflow Ops — gematik / eRezept integration handoffs, Krankenkasse-Abrechnung automation (KV-Abrechnung quarterly cycle, EBM / GOÄ code lookup support), post-visit comms (consent-gated reminders, NPS, follow-up scheduling), and DiGA activation chase sequences for the ~50% unredeemed-prescription gap. Areza does not write to clinical records; it orchestrates the surrounding comms and document flow, with explicit Art. 28 processor scope and EU data residency.

Knowledge Bot — trained on the operator's own patient FAQ, DiGA-Anleitung / Handbuch content, clinic protocols, aftercare instructions, and Kassen-coverage rules. Full HWG §3 disclaimer on every surface; no diagnostic claims, no clinical decision support, no triage. The bot routes any clinical question to the practitioner and answers operational ones only — coverage, pricing, opening hours, document checklists, what to bring.

Growth Stack — two distinct motions. B2B for hospital and ambulatory procurement runs long sales cycles, association-mediated via BVMed, BVITG, Bitkom Health, and ABDA where pharma-adjacent; the marketing surface is RFP-grade content, evidence indexes, and AI-search-friendly category pages. D2C for DiGA-app marketing under HWG ships compliant claim copy, schema-tagged review content, and Krankenkasse-listing pages keyed to coverage-name × Indikation pairs.

German digital-health search is fragmented across the BfArM DiGA-Verzeichnis, the 95 individual Krankenkasse content estates (AOK Gesundheitsmagazin, TK Techniker, Barmer Magazin), Apotheken-Umschau, Netdoktor, Onmeda, jameda, gesund.de, IhreApotheken.de, and clinic-specific pages.

No single source dominates a typical buyer journey from symptom to DiGA to clinician to prescription. AI Overviews routinely surface English-language US sources for German queries and miss the BfArM listing, the Krankenkasse coverage rules, and the local clinic — precisely the places a German patient or referring clinician needs to land.

The health vertical sits near the top of AI Overview share globally at ~43%, and the German rollout on 26 March 2025 means the citation surface is being rebuilt in real time.

The wedge for German operators with E-E-A-T-strong content tied to a real clinical or service entity is wide open: native-DE Foundation pages with BfArM IDs and MDR statements where they apply, AI Search retainer content keyed to coverage-name × Indikation queries, and a Knowledge Bot under HWG disclaimer that handles the operational questions clinic staff currently absorb by phone.

Voice is the other side of the gap. Twenty-five to 40% of German clinic calls are missed during peak hours; Doctolib has signalled the category's reimbursability by acquiring Aaron.ai. Areza's wedge is the strict-scope voice + structured-content + workflow layer that pairs cleanly with whatever clinical AI the practice has already procured under MDR — never replacing it, always adjacent to it.