PharmaLineAI is designed for the messy reality of pharmacy calls: it can answer routine questions, but it must verify identity before patient-specific detail and move clinical or sensitive queries to humans.
The agent must match surname, date of birth and postcode before returning prescription details. Failed verification returns no patient data, and repeated failures on a call stop further attempts and pass the caller to the team.
The public offer is routine call deflection, not clinical advice. Sensitive intents are routed to a human pharmacy team.
Significant access and mutations are recorded in an append-only trail, including patient IDs as plain columns so the trail survives record deletion.
The project is designed for UK/EU hosting and processor review before live pilots process real patient data.
The current product workflows are implemented locally. A real deployment still needs external operational, legal and integration checks before processing live patient calls.
Telephony provider integration and call-recording policy
Processor/sub-processor review and privacy notice confirmation
DPIA, retention settings and pharmacy owner sign-off
Hosted deployment in a UK/EU data-residency region
Live fallback and escalation runbook for staff
We can walk through routine intents, identity checks, audit events, escalation boundaries and deployment assumptions before any live caller is routed through the agent.
Book a trust review