HLC Hairline Clinic

What we did

1. 01

   Technical clean-up

   We started with a full audit: 182 URLs, 27 critical findings, more than 80 fixes. Among them: a backdoor planted in the code, database backups sitting inside the web directory, and third-party access that we revoked. Then came a staging environment, so that no change goes live unchecked, and clean caching.

   Result: a website the clinic controls again, with a review step before every change.

2. 02

   Optimisation for AI systems

   Search engines read HTML. AI assistants need more: structured data on the clinic, its doctors and treatments, clear signals about what may be crawled, and content that can be quoted as an answer. We added medical schema to the clinic and article pages, set content signals in robots.txt, and put in place an API catalog and markdown delivery for agents.

   Result: the agent-readiness score in our audit rose from 25 to 58 out of 100. Since June 2026 Google Analytics reports a separate "AI Assistant" channel: around 1,200 sessions from ChatGPT between June and August, a channel that did not exist before.

3. 03

   Four languages that work

   The site existed in English, German, French and Spanish, but the translations were barely reachable for search engines: wrong language references, a ghost entry for Italian, footer links pointing at internal IDs instead of pages, duplicate forms, a results sitemap from 2023. We cleaned up the language structure and checked each language version individually.

   Result: sessions in German went from 61 to 879 per month (March versus July 2026), French x5, Spanish x4.5.

4. 04

   Blog rebuilt

   We took twelve outdated posts out of the index instead of carrying them along. Three new specialist articles were written in four languages each, every one approved by a physician, plus a cost cluster of eight pages. Since August every text goes through a fixed review procedure before it is published.

   Result: fewer pages, but pages that answer a concrete question and get found through search and AI assistants.

5. 05

   Chatbot on the website

   On this foundation we built a chatbot that answers common questions from an approved knowledge base and leads to the consultation request. It runs on our own infrastructure, not at a SaaS provider, and gives no medical assessments.

   Status: in trial on the clinic's staging environment, not yet public.

6. 06

   Clinic CRM for lead handling

   HLC works with a CRM built for lead handling in clinics rather than a generic sales CRM. We extended it internally: a maintained knowledge base with prices, procedures and rules for managing expectations delivers draft replies directly in the ticket, instead of every reply being written from scratch. For the day of surgery the system summarises the entire enquiry history for the medical team.

   Status: Approval of every reply stays with a person.

Numbers

Source: Google Analytics 4, the clinic's property, time zone Europe/Zurich. March 2026 is the last full month before the project started on 12 April 2026. Sessions are not people. The numbers describe the project period; other influences such as seasonality cannot be ruled out completely.

Next step

If your clinic website shows similar symptoms, the audit is where to start. It shows whether technology, visibility in AI systems or lead handling needs attention first.