Clinics · Appointments
MixedOutpatient clinicsAppointment calls and reminders for clinics, with staff on anything clinical
Front desks lose hours to booking calls and reminder rounds, and no-shows still slip through. Here is how the parts fit: AI voice for bookings and confirmations, workflows for WhatsApp and SMS, and staff for anything clinical.
Valnora & Elric team5 min read
- WorkflowsThe visual builder: blocks, models and integrations wired into agent logic.
- TablesA database, built in, for cases, queues and decisions.
- DashboardsCharts, tables and KPI tiles over governed metrics.
- MonitorsWatch a metric and alert the right people when it moves.
- MetricsNamed, reusable measures every chart and agent reads the same way.
- LogsEvery run traced block by block.
A busy outpatient clinic spends a large part of the front desk’s day on the phone. Patients call to book, move or cancel. Staff ring round the next day’s list to confirm. Some patients still do not turn up, and the slot is lost because nobody knew in time.
Most of that work is scheduling, not care. This post shows how a clinic can hand the scheduling to AI voice and workflows, keep every clinical question with staff, and see no-shows on one dashboard.
What the setup covers, and what it never does
It covers three jobs: answering booking calls, reminding and confirming patients before the visit, and tracking who did not come.
It never gives clinical advice. Questions about symptoms, medicines, test results or whether to come in at all are passed to a person. That line is written into the call instructions and into the workflows, and it is the first thing to agree with your clinical lead.
How the parts fit
From a booking call to a confirmed appointment
Calls · Agent Inbound & Outbound
- 01Patient callsPatients call to book, move or cancel an appointment.
- 02AI voice agentAnswers instead of an IVR menu, finds a free slot and books it in the clinic’s system. Places reminder and confirmation calls.
- 03Staff handoffSymptoms, medication or results questions are passed to the front desk or a clinician, never answered by AI.
Workflow · Schedule and reminders
- 04Appointments tableOne row per booking with status, or the clinic system reached through the API block.
- 05Reminder workflowRuns on a schedule, picks tomorrow’s bookings and sends a WhatsApp template or a Twilio SMS.
- 06Reply receivedThe WhatsApp webhook trigger or an SMS-received trigger starts a workflow that updates the booking status.
Dashboard · No-shows
- 07No-show dashboardNo-show rate by day, department and reminder channel, read from the table.
- 08Unconfirmed monitorAlerts the front desk by email, in-app or Slack when too many of tomorrow’s slots are unconfirmed.
Calls. Agent Inbound & Outbound is the Valnora & Elric product that answers and places calls with an AI voice. It is available by invitation. For a clinic, it answers instead of an IVR menu, finds a free slot that matches the booking rules, and books the appointment in the clinic’s system. The day before, it can place reminder and confirmation calls to patients who prefer a call. When a caller raises something clinical, it creates a support query for the front desk instead of answering.
The schedule. Bookings live in an Agent Serve table: one row per appointment with patient reference, department, time and status. If the clinic already runs a practice management system with an API, workflows read and update it through the generic API block instead, with the key stored as a workspace secret.
Reminders. A workflow built in the workflow builder runs on a schedule, picks tomorrow’s bookings and sends each patient a WhatsApp template with reply buttons, or an SMS through Twilio. Replies come back through the WhatsApp webhook trigger or the Twilio SMS-received trigger, and a second workflow sets the booking to confirmed, cancelled or "wants to reschedule". Patients who want a new time get a call back from staff or from the voice agent.
What patients and staff see
What the solution produces
WhatsApp reminder with reply buttons (illustrative)
MessageHello Priya, this is a reminder of your appointment at City Outpatient Clinic tomorrow, 10:30, with the physiotherapy team.
Please reply with a button below.
[Confirm] [Reschedule] [Cancel]
Staff handoff from a booking call (illustrative)
Record- Caller
- Existing patient, identity confirmed by date of birth
- Booked
- Follow-up, Thursday 16:00, general medicine
- Not answered by AI
- Asked whether to keep taking a prescribed medicine before the visit. Told a member of staff will call back.
- Handed to
- Front desk callback queue, flagged for the duty nurse
The reminder is short, uses an approved template and names no condition or treatment. The handoff record shows what the AI booked and, just as clearly, what it did not answer and who now owns it.
Tracking no-shows
Appointments · next 7 days (illustrative)
The no-show dashboard, built with Dashboards, reads the appointments table. Because the reply workflows write each confirmation and cancellation as it arrives, the panels update as rows land. Useful views are no-show rate by week, by department and by reminder channel, and a list of tomorrow’s unconfirmed slots.
A monitor re-checks the unconfirmed count as often as every five minutes and alerts the front desk by email, in the app or in Slack when it passes a level you set. That gives staff time to call the remaining patients or offer the slot to the waiting list. All figures in the example above are illustrative.
Consent and health-data rules
Reminders are personal data, and in a clinic they are often health data. Before going live:
- Record consent per channel. Only send WhatsApp or SMS reminders to patients who agreed to them, and keep that flag on the appointments row so the workflow can skip anyone without it.
- Say less in messages. Use the time, place and team name. Leave out diagnoses, procedures and test names.
- Check the law where you operate. HIPAA in the US and the DPDP Act in India are examples of rules that apply to patient contact and health data. Agent Serve does not claim compliance with either; ask your counsel what they require of your setup.
Controls and governance
Run logs
Every workflow run records each block’s inputs, outputs and errors, so a reviewer can trace how a result was reached.
Audit trail
Proposals, approvals, declines and executed actions are written to the workspace audit log, which admins can export.
PII detection
A Guardrails block detects personal data and blocks or masks it before text reaches the next step.
Roles and permission groups
Admin, write and read roles per workspace, plus permission groups that limit which models, blocks and integrations a team can use.
Row-level access
Row-level policies narrow what each person sees in tables, dashboards, exports and scheduled reports.
Deployed versions
Workflows run from a deployed version, so a change is reviewed before it reaches production runs.
Self-hosting
Run Agent Serve in your own cloud or data centre, next to the systems it reads.
More on how Agent Serve handles access, data and audit on the security page.
Every workflow run is logged block by block, and PII redaction in Settings → Data retention can mask personal data in logs, workflow inputs and block outputs. Row-level access keeps each department to its own bookings. Clinics that need patient data to stay on their own infrastructure can run Agent Serve self-hosted.
Where to start
Start with reminders only, for one department, while staff keep answering the phones. Once replies flow into the table reliably, add the voice agent for booking calls, and review its handoff records with the clinical lead every week for the first month. For what happens when a patient calls to complain, see handling patient complaints.
Questions teams ask
Will the AI answer questions about symptoms or medicines?
No. The call instructions treat any question about symptoms, medication, test results or treatment as clinical. The caller is told a member of staff will help, and a support query goes to the front desk or the duty clinician. The AI handles scheduling only.
Do we have to move our schedule into Agent Serve?
No. A table in Agent Serve is the simplest start. If your practice management system has an API, workflows can read and update bookings there with the generic API block, using a key stored as a workspace secret.
Is Agent Inbound & Outbound part of Agent Serve?
It is a separate Valnora & Elric product, available by invitation. It handles the calls; Agent Serve holds the workflows, tables and dashboards around them. Bookings it makes can be exported or synced into a table during setup.
What you need
- An invitation to Agent Inbound & Outbound, and the clinic’s booking rules: slot lengths, departments, who may book what
- An appointments table in Agent Serve, or API access to your practice management system for the API block
- A WhatsApp Business account with approved reminder templates, or a Twilio number for SMS
- A record of each patient’s consent to be contacted, and on which channel
- A written list of what counts as clinical and must go to staff
- Advice from counsel on the health-data rules where you operate