Patient access & scheduling
Make patient access feel available, not automated
Patient access fails quietly: the hold queue at 9am, the voicemail nobody returns, the preparation instruction that never arrived so the appointment is wasted. This blueprint gives every patient a consistent front door for scheduling, preparation, and administrative questions, with clinical judgement and urgent presentations routed to qualified staff on the first turn.
- The team
- 3 AI worker roles escalating into 2 human decision owners.
- The process
- 6 orchestrated stages, each with a named decision owner.
- The channels
- Voice, Web chat, SMS, WhatsApp on one shared context.
The shift
What breaks today, and what changes
The challenge
- Patients wait on hold for administrative tasks while staff move between the scheduling system, the record, and the phone.
- Missed calls become missed appointments, and unused slots cannot be refilled fast enough to matter.
- Preparation instructions are delivered inconsistently, so appointments are wasted on unprepared patients.
What changes
- Patients reach a responsive front door on voice or chat, including outside clinic hours.
- Urgent and clinical language routes to qualified staff immediately, with the full conversation attached.
- Confirmations, preparation instructions, and reminders go out consistently on the channel the patient uses.
The team
Every seat has one job and a defined level of autonomy
This is what gets configured in Team Canvas: role-scoped AI workers, the humans they escalate into, and a trust level on each connection that sets what a worker may do on its own. Trust starts tight and widens on evidence.
AI workers
Access worker
Answers administrative and preparation questions from approved, grounded content.
- Owns
- Non-clinical guidance
- Trust level
- Acts within policy — approved knowledge only, no clinical content
Scheduling worker
Finds real availability by appointment type and confirms, moves, or releases the slot.
- Owns
- The appointment journey
- Trust level
- Acts within policy — writes only to permitted appointment types
Safety router
Watches every turn for urgent and clinical language and hands off the moment it appears.
- Owns
- Urgent-intent detection
- Trust level
- Acts within policy — escalation is mandatory, never discretionary
Human decision owners
Clinical staff
Takes every clinical question, symptom description, and urgent presentation.
Owns: All clinical judgement
Access supervisor
Reviews scored conversations and every escalation trigger for correctness.
Owns: Safety review and policy tuning
The process
The conversation starts the work. The workflow finishes it.
Each stage records the decision that was made and the role accountable for it — which is what makes the run reviewable afterwards rather than a black box. Where the platform executes a stage rather than a person or agent performing it, that is named too, and accountability still sits with the role.
- 1
Understand the need
Is this administrative, clinical, or urgent? Anything but administrative leaves the flow.
Access worker
- 2
Verify
Does this specific action require identity verification before it proceeds?
Access supervisor
Run by workflow
- 3
Answer or schedule
Is there approved content for this question, or a real slot for this appointment type?
Scheduling worker
- 4
Safety check
Has any turn contained urgent or clinical language? If so, hand over now.
Safety router
- 5
Confirm & prepare
Which confirmation, preparation instruction, and reminder does this appointment need?
Scheduling worker
Run by workflow
- 6
Recover
Cancelled or missed — offer the slot onward and rebook the patient?
Scheduling worker
Run by workflow
Decision rules
The non-negotiables encoded in the workflow, not left to a prompt.
- Urgent intent immediately follows the configured emergency path and ends self-service.
- Clinical questions transfer to qualified staff — no worker offers symptom interpretation or advice.
- Only the minimum record needed for the requested task is exposed to any worker.
- Identity verification is required before any appointment or record detail is discussed.
Controls & guardrails
What makes this safe to run in production, and provable afterwards.
- Role-limited knowledge with clinical content excluded from worker scope.
- Mandatory urgent-intent routing that cannot be overridden by conversation flow.
- Least-privilege record access scoped to the task in hand.
- Full audit trail across bookings, cancellations, and every clinical handoff.
Platform capabilities
What this blueprint runs on
Nothing here is bespoke. Each blueprint is a configuration of the same platform, which is why the second one you launch reuses the governance you already reviewed.
Channels
- Voice
- Web chat
- SMS
Systems it reads and writes
- Scheduling system
- Patient administration system
- Reminder & messaging gateway
- CRM
Measurement
What we instrument from day one
These are the measurements the pilot puts in place, not benchmark results. You set the targets against your own baseline — and the same instrumentation is what decides whether a worker's trust level widens.
- Access rate
- Measures how many access attempts reach a resolution instead of a hold queue or an unreturned voicemail.
- Slot utilisation
- Tracks how much released capacity gets refilled, and how quickly, once cancellations are handled conversationally.
- Escalation precision
- Reviews every clinical or urgent handoff to confirm the trigger fired on the correct turn — the safety metric that gates trust expansion.
Answered vs abandoned
Booked vs available
Clinical routing
Rollout
How this one goes live
A deliberately narrow start, supervised, with autonomy widened per role once the record supports it.
Weeks 1–2
Administrative questions only
Start with directions, hours, preparation, and general service questions. No scheduling writes yet — this stage proves grounding and escalation accuracy.
Weeks 3–4
Add scheduling for one appointment type
Enable booking and rescheduling for a single low-risk appointment type, with confirmations and reminders on the patient’s channel.
Week 5+
Extend types and add recovery
Widen appointment types and switch on cancellation recovery so released capacity is offered onward automatically.
Questions we get asked
Patient access & scheduling, in practice
- Can an AI worker give patients clinical advice?
- No. Clinical content is excluded from the workers’ knowledge scope, and a dedicated safety router escalates clinical or urgent language to qualified staff on the turn it appears. The workers handle scheduling, preparation, and administrative questions from approved content only.
- How is patient data protected?
- Access is least-privilege: each worker sees only the minimum record needed for the task it is performing, identity verification gates any disclosure of appointment or record detail, and every access and action is written to a tamper-evident audit trail.
- What happens to a call that arrives after hours?
- The same team answers, because the blueprint is not staffed by shift. Administrative requests are resolved immediately; anything clinical or urgent follows the configured out-of-hours escalation path rather than waiting for the next working day.
What changes in Healthcare
Clinical and urgent language routes to qualified staff on the first turn, and workers see only the minimum record needed for the task.
Read the full Healthcare analysisRelated blueprints
Teams that run next to this one
Next step
Build this team around one measurable win.
We map your version of this process, compose the worker roles, connect your systems, and agree the escalation points with the people who own them — then run it supervised.