A Swiss software company selling an AI chatbot into healthcare needed conversations with clinics. Compliance-safe cold email with tight segmentation got nearly 1 in 11 prospects to reply.
Swiss software company with an AI chatbot for healthcare providers.
Regulated market, cautious buyers, no outbound pipeline.
Segmented cold email with compliance guardrails and DNC enforcement baked in.
8.95% reply rate overall, 28% on the best segment, zero do-not-contact violations.
Context
A Swiss software company that builds a proprietary AI chatbot platform for healthcare. The product is Swiss-hosted, GDPR and Swiss data sovereignty compliant, features a multi-agent architecture with built-in CMS and live-chat handoff, and can go live in roughly three days. Existing customers include major cantonal hospitals, national insurers, and public-sector health organizations across Switzerland.
They had strong inbound traction with large institutions but no systematic outbound function. The addressable market, German-speaking Swiss healthcare organizations with 20+ employees, is small by global standards, and every missed opportunity or compliance mistake would be visible across the entire sector. They needed cold outreach that was precise, culturally fluent, and compliant enough to survive scrutiny from hospital IT directors and compliance officers.
Constraints
Swiss German healthcare is one of the hardest cold email markets on the planet, not because of volume, but because of constraints that stack on top of each other. The real constraint was never "how do we get more leads," it was "how do we reach the right 200 organizations without a single mistake."
Architecture
I built a three-layer system: precision discovery, bulletproof DNC compliance, and culturally adapted copy. Data flows left to right, from a hard-ceiling target universe through segmentation and compliance rails, out to segmented campaigns, and back as replies.
I used lookalike discovery to find healthcare organizations similar to the client's best existing customers. Three batches produced ~273 net new companies, of which 212 passed ICP qualification. I filtered on geography (German-speaking cantons), sector (hospitals, clinics, rehab centers, nursing homes, health insurers), size (20+ employees), and existing chatbot detection, with encoding-safe regex to handle Swiss location data quirks that defeated standard keyword matching.
Once the core universe was worked, I expanded into 5 adjacent verticals: nursing homes, psychiatric clinics, home care organizations, and health insurers, adding 320+ new companies while keeping the same qualification bar.
Campaign 1 left compliance debt: contacts from organizations that said "Not Interested" reappeared in Campaign 2, and the project nearly got cancelled. I built a four-layer DNC system: a DNC_Companies table (219 organizations), a DNC_Contacts table (1,462 contacts), automated SmartLead removal, and a boolean Push_Ready gate. The result was zero DNC leaks across Campaigns 2 and 3.
I built separate Du and Sie variants with a Sie-option PS in every sequence, replaced "Hoi" with "Hallo" because the casual greeting read as too informal for hospital directors, and used gender-aware German salutations from enrichment data. Round-robin distribution across 3-day windows meant no two people at the same hospital got emails on the same day, across ~60 sender variants for deliverability protection.
Results
| Metric | Result |
|---|---|
| Aggregate reply rate (Campaign 1) | 8.95% (340 / 3,796) |
| Best sub-campaign reply rate | 28% |
| Interested prospects (all campaigns) | 49+ |
| DNC companies tracked | 219 |
| DNC contacts tracked | 1,462 |
| DNC leaks after system build | 0 |
| Sender variants in rotation | ~60 |
| Total addressable market (net new cos) | ~212 |
| Adjacent verticals discovered | 5 |
Want numbers like these for your pipeline?
Book a 15-min callFit
Stack





See how a compliant, culturally-adapted outbound system gets built for your market, without a single do-not-contact mistake.
Book Your Free Strategy Call