Set the rules once.
Nothing slips after that.
Every aspect of AI behaviour, what it asks, captures, escalates, and how it responds, configured at the site level. Then the same rules run the follow-up your front desk never reaches: rebooking no-shows, chasing unbooked referrals, refilling cancelled slots.
- Saturday hours
- Referring-GP fast escalation
- After-hours web booking
- MRI pricing read-back
A site isn't a category.
A radiology group with five locations doesn't run them identically. Saturday hours differ. The referring-GP escalation rule differs. The "what's the cost of an MRI" answer differs. TriageWorkflow lets the operator set each one, per site, in a UI that's quiet enough to disappear behind the data.
The same rules then govern what happens after the call. Inbound gets answered because the phone is ringing; outbound doesn't. Configuration and follow-up are one surface here, because in practice they are one decision.
Monochrome by design. No coloured KPI tiles competing for attention. Just the work.
The operator's
quiet workspace.
Per-site configuration
Different rules for different sites in the same group. What the AI asks, captures, and escalates, set per site.
Real-time analytics
Calls, chats, bookings, escalations, live counts and trends. Drill from a number to the underlying transcript.
Custom escalation rules
Route referring GPs differently from patients. Set the vocabulary that counts as an emergency, though not what happens next: that is always the 000 instruction.
Radiologist rostering (v1.01)
Specialist attendance and on-call schedules drive what TriageVoice / TriageChat say is available, in real time.
Exportable reports
CSV / PDF for management. KPIs aligned to your existing operating cadence, daily, weekly, monthly.
Audit log
Every change to AI behaviour, configuration, or escalation rule, captured, attributed, time-stamped.
- ↻No-show · MRI kneeNorthgate · rebooked Thu 9:15Rebooked
- →Unbooked referralRiverside · 2nd attemptContacting
- ✦Cancelled slotLakeside · waitlist filledFilled
- ·Prep reminderNorthgate · CT abdomenQueued
The revenue you lose is the work no one had time to do.
The no-show that needs rebooking, the referral sitting unactioned, the cancelled MRI slot nobody filled. It's not that the front desk won't; it's that there's never a quiet hour to. So it slips, and every slip is a scan that didn't happen.
TriageWorkflow runs that follow-up continuously, in the background, on the rules you already set: turning the work that used to fall through into bookings on the schedule.
Follow-up that
never runs out of time.
No-show recovery
Every missed appointment is contacted automatically and offered the next suitable slot, the rebooking call your front desk never has time to make.
Referral chasing
Unbooked referrals are followed up until they convert or close, so a faxed request doesn't quietly become lost revenue.
Prep & appointment reminders
Timed SMS or voice reminders with prep instructions, fewer arrivals turned away for the wrong prep, fewer no-shows in the first place.
Waitlist backfill
When a slot is cancelled, the waitlist is worked in order and the gap is filled before it becomes an idle scanner hour.
Rule-governed outreach
Every outbound action runs on the same per-site rules as voice and chat, contact windows, frequency caps, opt-outs, and escalation all honoured.
Full audit trail
Who was contacted, when, on which channel, and what they said: every action logged and attributable, to the same trail as the rest of the platform.
What operations managers ask.
What outbound work does TriageWorkflow do?+
Can rules differ between sites in the same group?+
Is outbound contact governed by the same rules as inbound?+
Can we see what the agents did and who changed what?+
Does it know the radiologist roster?+
From the blog
What we are learning, written down.
- What can AI agents actually do in radiology?Not AI reading scans. What an AI agent actually does across reception, booking, referrals, capacity and follow-up, and what it must never touch.
- We called 100 radiology practices after hours. Here is what answered.Original research: the same 100 radiology practices, rung on Sunday evening. Not one reached a person, and eleven were answered by an AI agent.
- We called 100 radiology practices. Here is what we found.Original research: we rang 100 Australian radiology practices and timed every call. Eight reached a person straight away; the median wait was 90 seconds.
Book a radiology network assessment.
We'll map it to your network.
30 minutes with a solutions engineer. We model your call volume, centre count, and integrations, then show the AI agents running on your own scenarios, and answer everything compliance- and security-related upfront.