Updated 2026-10-03 · 2 min read
Many clinics have tried a chatbot at some point: a menu with «book an appointment», «opening hours» and «prices» buttons. It works as long as the patient asks exactly what the menu expects. As soon as they write «have you got anything Thursday afternoon with Marta? it's for my knee», the menu doesn't know what to do.
An AI assistant doesn't work with menus: it understands what the patient writes, checks the calendar and replies.
The difference, in practice
| Menu chatbot | AI assistant | |
|---|---|---|
| How you write to it | Pressing options | As you would to a person |
| Calendar | A form or a link | Checks and books within the conversation |
| Changes and cancellations | Usually by phone | By message, with the real calendar |
| Voice notes | Doesn't understand them | Transcribes and replies to them |
| Cases out of the ordinary | Gets stuck | Passes the conversation to the team |
What neither of them should do
Neither a chatbot nor an AI assistant should assess what a patient describes. Sorting isn't diagnosing: the right thing is to note the reason in the patient's own words, offer the first available appointment with the physiotherapist and, in the face of a red flag, point them to their doctor without booking.
A well-built assistant has it written into its rules; it doesn't leave it to the judgement of each conversation.
How to choose
If the clinic receives few messages and they're all the same, a menu may be enough. If patients write the way they speak, send voice notes, change sessions and ask about packs, an AI assistant saves hours of reception work every week.
ClaudIA is an AI assistant: it handles WhatsApp, Instagram and, in Spain, the phone; it books with the real calendar; it keeps track of packs; and it applies the clinical red line in every message.
Questions clinics ask
Is an AI assistant the same as a chatbot? No. A menu chatbot responds to predefined options; an AI assistant understands what the patient writes, checks the calendar and books within the same conversation.
Can the AI assistant assess symptoms? It shouldn't. ClaudIA notes the reason in the patient's own words, doesn't assess it and, in the face of a red flag, tells them to see their doctor without booking.
What happens if the assistant doesn't understand something? It asks, and if the case is out of the ordinary, it passes the conversation to the team with an alert.