The business and operational need
Anastasakis Hair Clinic already had a detailed six-month post-operative protocol. The problem was not a lack of knowledge. It was how that knowledge reached the patient.
A static handout asks patients to work out which instructions apply today, remember medication times and distinguish questions covered by the protocol from those that require contacting the clinic.
The solution
We designed a Progressive Web App accessed through a personal link, with no App Store required. Based on the procedure type and date, it displays the relevant instructions, medication, videos and restrictions each day.
Personalised daily schedule
Instructions are generated deterministically from the clinic’s approved protocol. They are not improvised by AI.
Reminders and essential actions
The system organises medication and actions by day and time, so patients see what they need now, rather than the entire six-month plan at once.
AI assistant with strict guardrails
The assistant answers only from approved medical content. When a question is out of scope or a potentially critical symptom appears, it does not improvise or offer reassurance. It directs the patient to the clinic.
Admin and CRM sync
The clinic’s team can create access, manage content and connect the application to its existing CRM workflow.
Where AI was used, and where it was not
The guiding principle was simple: safety comes before spectacle.
- The daily protocol and medication rules are deterministic.
- AI is used to understand the question and explain approved content.
- Questions beyond the knowledge base and critical signs are escalated to a person.
- AI activation is subject to the required legal and medical approvals.
The infrastructure patients do not see
Behind the simple experience are rules for each phase and procedure type, edge-case handling, access permissions, notifications, CRM integration and data management within the European Union.
Accounts and data were set up in the clinic’s name. The architecture makes the system an asset the clinic owns, not rented access to a closed platform.
What we can say with confidence today
The system has been delivered and operates as a real client application. It supports personalised protocols, reminders, CRM sync, admin functions and a controlled AI assistant.
Adoption, question volumes and any reduction in calls need an adequate measurement period and evidence from the clinic. We do not present them as results before reliable data exists.
What this demonstrates for other projects
Not that every process needs a healthcare app. It shows that a complex protocol can become a safe, personalised and measurable experience when rules, AI and human responsibility are designed together.