Case Study: AI Assistant for a Therapy Practice
A small occupational therapy practice needed less manual coordination around inquiries, waiting list work, documentation, patient handouts, and daily organization. The goal was not a chatbot demo. It was a practical assistant layer that fit the operating rhythm of a real practice.
Starting point
The practice had the typical pressure pattern of many small service businesses: patient inquiries arrived through different channels, documentation competed with therapy time, and the owner carried too much organizational load personally. None of these tasks justified a full-time administrative hire on their own. Together, they created a constant drag on the business.
The system was built around the work that actually repeated: website inquiries, waiting list follow-up, therapy documentation from audio or transcript material, patient-friendly PDF handouts, daily task lists, and recurring practice administration.
What was implemented
Website inquiry flow and waiting list support
The website form collects the relevant information before a human follows up: contact details, therapy area, urgency, and the reason for the request. That turns scattered inbound messages into a structured queue.
Documentation support from voice and transcript material
Therapy notes can be turned into structured drafts and PDFs. The therapist remains responsible for review and final judgment, but the first structured version no longer has to be written from scratch.
Patient handouts and reusable materials
The assistant helps draft patient-friendly exercise and information sheets, then turns them into reusable PDF materials for recurring therapy situations.
Daily organization
Open tasks, reminders, shopping lists, documentation backlog, and waiting list items are kept visible. The value is partly time saved and partly fewer things falling through the cracks.
Measured activity by 16 June 2026
The analysis is based on anonymized local working files, memory logs, website request sync data, and generated documentation or handout artifacts. Patient names are not used.
Estimated business effect
A strict count of assistant interactions would understate the result. The business value came from several small workflows working together: fewer manual inquiry steps, faster documentation drafts, reusable handouts, cleaner waiting list handling, and less daily re-sorting of open work.
For a small practice, the most useful number is not only the total hours. It is where the hours were removed: after-hours documentation, repetitive inquiry handling, and the mental load of keeping every open item in the owner's head.
What made the project work
- The assistant was built around existing practice work, not around a generic AI feature list.
- The workflow kept professional judgment with the therapist. AI produced drafts, structure, and reminders.
- The system covered several small bottlenecks instead of betting everything on one large automation.
- Measurement stayed conservative. Only numbers with a clear source were used for the public case study.
Client testimonial
"The assistant takes work out of the practice that used to sit in my head or land after hours: inquiries, waiting list follow-up, documentation drafts, handouts, and daily open tasks. It does not replace therapeutic judgment. It gives me structure, prepares the repetitive parts, and helps me keep the practice moving without losing track of details."
Elisabeth Aster, occupational therapy practice owner
When this pattern fits
This pattern fits service businesses where the owner is still the operational bottleneck: therapy practices, trades businesses, local professional services, and small teams with repeated documentation or customer communication. It is less useful when the work is too irregular, the owner does not want to change the process, or the business only wants a cheap chatbot on the website.