for Medmate
Built around what Medmate patients actually message about — where's my delivery, lost script SMS, can I book a doctor today, refund this consult, medical certificate — the agent answers and actions, without ever giving clinical advice.
Loaded with dummy data for a fictional patient — Emma Lawson, Melbourne, with a recent delivery and active scripts.
Click any prompt to copy it.
AI support that understands online consults, e-prescriptions, partner pharmacy delivery, medical certificates, and Australian healthcare regulation.
Pulls live shipping status from the courier, flags late deliveries, and opens an inquiry or recall if something's wrong — without reading URLs aloud on voice.
Quotes back the booking (doctor, time, $49 charge) before booking. Never describes what the doctor "will" prescribe — that's the consult's job.
Looks up active scripts, resends QR codes when patients lose the SMS, and explains repeat timing — at any AU pharmacy, not just the Medmate partner.
Starts the certificate request, quotes back days and cost, and triages routing — text-based for 1-2 days, doctor video for longer or backdated.
Honest about Medmate's auto-approve vs manual-review policy. Doesn't promise refunds it can't deliver. Files clinical-outcome disputes to a senior reviewer.
If a patient discloses self-harm or suicidal ideation, the agent surfaces Lifeline, Beyond Blue, and 000 calmly — and stays out of clinical advice the rest of the time.
Wired up specifically for Medmate's Australian patient base.
Pulled the live medmate.com.au FAQ into the knowledge base and grounded answers in your real product terms.
Seven topic-matched workflows — order tracking, consults, e-scripts, medical certificates, refunds, plus a small-talk intro and open-conversation fallback.
Ten mock tools — patient profile, orders, tracking, active scripts, plus write actions to book consults, resend scripts, request refunds, and issue certificates.
Three STEER guardrails — no clinical advice, AU crisis-line surfacing, and tight Medmate scope — plus voice rules and identity-gate pattern before sensitive lookups.
From this demo to production in weeks, not months.
Replace dummy data with real patient records, live consult history, and the real Medmate product catalogue. Tighten the agent against real patient conversations.
Order tracking, script lookups, consult status — the agent answers and diagnoses without touching anything in production.
Book consults, reschedule, refund unfulfilled orders, resend scripts, request certificates. Enable progressively once you've validated read-only.
Add voice and SMS. Same workflows, same brand voice, every touchpoint patients reach you on — chat widget, support email, AU phone, SMS.
Lorikeet is built for businesses handling sensitive patient data — AU privacy and data residency are non-negotiable.
Multi-region deployments. Patient data stays in AU. Pick the jurisdiction that suits your privacy posture.
Phased API access — start read-only, enable write-back when you're ready. The agent only sees and actions what you've explicitly authorised.
Security docs, sub-processor list, vendor questionnaire, and DPIA available on request — we can share access with your compliance reviewer.
"Where's my order?" is the #1 ticket. The agent surfaces tracking in one turn — including ATL status and the realistic estimate from the courier.
The agent handles the top 60-70% of patient questions — bookings, script resends, delivery tracking, certs — so your team focuses on the clinical or complex.
Patients get instant answers at 11pm Sunday. Medical certificates and consult bookings still route to real doctors — the agent just handles the front door.
The agent never diagnoses, recommends medications, or evaluates symptoms. Built for healthcare, by design.
Deploy with the Medmate brand and voice. The agent represents your product, not a third-party chatbot.
Handle 3x the patients without growing the support team. The AI grows with your patient base.
Happy to walk through how this works in production and answer any questions.
Talk to the team