Inquiries outside practice hours
Patients call on Sunday evening because they're in pain and need an appointment Monday morning. No one picks up. By Monday morning, the person is with the competition.
[ Case study · physiotherapy ]
How I built a digital reception for my wife's physio practice in Hinwil and Bauma — and what you can learn from it for your own practice.
Full transparency: PhysioHeld is my wife Franziska's practice. That's exactly why it's a good showcase — I saw firsthand what a practice actually needs, and could iterate quickly and honestly.
[ What was before ]
PhysioHeld already had a functioning practice in two locations — Hinwil and Bauma — with a solid patient base and full schedules. What was missing wasn't a marketing boost. It was three very concrete gaps in daily operations:
Patients call on Sunday evening because they're in pain and need an appointment Monday morning. No one picks up. By Monday morning, the person is with the competition.
"Do I need a referral?" — "Does insurance cover this?" — "Where can I park?" Thirty times a day, interrupting real treatments.
The old website was functional but not built for patient acquisition — no clear calls to action, no SEO optimization, no visibility on Google for local search.
[ The solution ]
Instead of a marketing showpiece, I built three pragmatic layers — from "must" to "adds value" to "makes the difference".
Astro-based, fast practice site with clear location separation (Hinwil and Bauma), treatment offerings, team page, online booking request, and all legal requirements (Swiss DPA-compliant, Swiss hosting). Lighthouse performance > 95.
Embeddable AI chatbot in the bottom-right of every page. Patient questions are answered 24/7 — at night, on Sundays. The chatbot uses only the approved practice content (FAQ, treatment descriptions) and forwards structured inquiries directly to the practice. No free-form LLM babble, no hallucinations — the output guard checks every answer before sending.
Classic search engine optimization for "Physiotherapie Hinwil", "Physio Bauma" and similar long tails. Plus preparation for AI crawlers (GEO): structured data, FAQ schema, llms.txt, clear knowledge blocks that ChatGPT and Perplexity can use as sources.
[ Under the hood ]
For PhysioHeld I use proven, performant tools — not an experimental stack but a production setup that's still maintainable in five years.
| // Area | // Stack |
|---|---|
| Frontend | Astro + Tailwind CSS |
| Hosting | Cloudflare Pages (Swiss edge) |
| Chatbot | HeldChat (Spring Boot backend, embedding-RAG pipeline, Swiss servers) |
| Data protection | Swiss DPA-compliant, data processing agreement, no transfer to the US |
| Analytics | Plausible (cookieless, EU-hosted) |
| SEO | Search Console, structured data, local schema |
[ What it brought ]
The site went live recently. Some effects are measurable immediately, others take time. Current status:
// Measurable now
// Under observation · 90 days
// Qualitative effect
This case study is updated as valid three- and six-month data becomes available.
[ Lessons learned ]
80% of patient questions are the same ten routine topics. A good chatbot answers those reliably — and honestly forwards the other 20% instead of guessing.
Swiss DPA-compliant setup, Swiss servers, processing agreement — those aren't marketing arguments, they're knockout criteria. Without that foundation, no chatbot works in a practice.
Lighthouse score above 95 means: patients don't wait. On mobile, on the train, on weak networks — the site is there. That's measurably better for conversion and for Google rankings.
[ What this means for you ]
PhysioHeld is a model, not a special case. Most owner-run physio, occupational therapy, speech therapy, and family-doctor practices in German-speaking Switzerland have the same three problems — and for most, the same three-layer approach is worth it.
If you own a practice and you're reading this: let's talk for 30 minutes. I'll look at your current situation free of charge and tell you honestly whether a similar setup would work for you — or not.
[ See it yourself ]
The best case study is the real one. Look at the site, click through the chatbot, ask it a question. You'll see in 5 minutes what text could never show.
To the live site: physioheld.ch ↗