Human control
AI output is advice for the operator, never an autonomous commitment.
Trust
Confidence comes from visible boundaries.
CareSignal's direction is grounded in organisational control, traceable sources and the freedom to say “the evidence is not clear.”
Authorised evidence
Recommendations are grounded in approved sources, with provenance and uncertainty retained.
Tenant isolation
Calls, knowledge, configuration and audit records remain separated by tenant.
Data minimisation
Audio, transcripts, personal information and credentials are collected and retained only where necessary and approved.
Safe degradation
A failed connector or unclear source is shown as a limitation, not hidden behind confident language.
No invented proof
Customer names, outcomes, pricing and production-readiness claims require evidence and approval.
Questions about the public help centre?
Read how its anonymous session and knowledge boundary work.
Read the privacy guide