Service Context & Content
Two things tell a service’s AI what it needs to know: service context, a plain-language description of how the service runs, and the content library, the documents agents and AI encounters can draw on.
Service context
The Service context tab (the Agent context step in the setup wizard) holds free text that is given to every service agent run. Describe:
- how the service runs and the steps patients and staff go through
- who does what, and when
- when to escalate, and to whom
- how to use the content library, e.g. “Answer dietary questions from the ‘Post-op nutrition’ handout in the content library”
There is no graph or builder: it is saved as text with Save agent context. It replaces the old program Workflow text.
Drafting it with AI
Paste anything unstructured into Notes on how this service runs (a workflow write-up, an SOP, handover notes, escalation rules) and click Generate context with AI. Review and edit the draft before saving.
Other context agents receive
Service context is one layer. Agents also receive:
- each care team member’s agent context, built from their preferences (see Care Team)
- each patient’s context, built from their encounter notes (see Patients & Cohort)
Content library
The Content tab (the wizard’s Content library step) holds the clinical resources this service’s agents and AI encounters can use.
Adding a document
-
Upload a document: choose a PDF, give it a Title and an optional Description, and pick its type:
- Reference material: background the assistant can quote from, such as a handout or guideline.
- Counselling protocol: a structured script for a counselling or education session.
Click Upload & add to library.
-
Attach an existing resource: pick a document you have already uploaded elsewhere in Medbot.
A new document is processed before it can be used; until then it shows as pending or processing. Remove takes a document out of this service’s library.
Where the library is used
- AI encounters can have documents attached by default, attach the best matches automatically, and let the assistant search the library. See AI Encounters & Kiosk.
- Service agents with the Clinical resources tool can look things up in it.
- The patient agent answers health-education questions from it.