AI for patient messages, answered from inside the practice.
Patients write to their practice about everything: repeat scripts, results, forms, a worry that came up after the appointment. Reception staff answer most of it from templates and the record, and the rest goes to a GP. An Ermos unit drafts the replies on hardware inside the practice, so the messages, which are often frank about health, never leave it.
How it works on the unit
Emails and portal messages are indexed on the unit from the inboxes your practice already uses.
Routine replies draft in your wording: booking a review for a repeat request, explaining how results are given, sending a form.
Anything that needs clinical judgement is flagged for a GP rather than answered, with the relevant part of the record attached.
Every reply is checked and sent by a person through your normal systems.
The boundary, stated plainly
The unit drafts administrative replies. It does not prescribe, give results or give advice. Decisions about medication, including repeat prescriptions, are made by a GP in your clinical software.
Why on-premise matters here
Patients often write more freely than they speak: medication names, symptoms, family problems. On the unit those messages are drafted against inside the practice; pasted into a cloud tool, each one becomes a disclosure.
See what leaves the practice. What a repeat prescription request gives away in ChatGPT or Claude: marked up item by item.
Common questions
Can reception staff use AI to answer patient emails?
Does the unit approve repeat prescriptions?
What about messages that mention a third party?
See Dominion answering from your own documents.
A demonstration takes 30 minutes, uses no client data, and comes with no obligation.
Book a demo