The typical radiology clinic loses three out of every ten callers before anyone picks up the phone. Those abandoned calls represent missed scans, delayed diagnoses, and revenue walking out the door while staff scramble through Byzantine scheduling requirements that make booking a haircut look simple.
Avoca Systems thinks it has cracked that problem. The Y Combinator-backed Australian startup has deployed an AI scheduling platform across more than 250 clinics, where it fields patient calls around the clock and navigates the peculiar complexity of radiology bookings: which scanner is free, which technologist specializes in pediatric MRIs, whether the patient needs contrast dye, how many hours they must fast beforehand.
"About 99% of people who use the service don't even recognise that it's AI," said Sam Netherclift, who oversees the contact center at Jobfit, an occupational health network using Avoca's system. "They're happy with it, they just book."
That invisibility may be the point. Co-founders Max Shand and Campbell Mercer-Butcher designed Avoca to handle what generic scheduling software typically punts to a human operator. The platform processes inbound and outbound calls, parses faxed referrals, verifies insurance requirements, asks mandatory safety questions, and writes confirmed appointments directly into radiology information systems without manual data entry. It works nights, weekends, and holidays.
Shand brings a resume that suggests he knows something about scaling consumer-facing technology. He was the first employee at Afterpay, the Australian buy-now-pay-later juggernaut, and previously founded and sold another venture called Serenade. Mercer-Butcher spent years as a principal engineer at Honey Insurance, Macquarie Bank, and property management platform Ailo. Their eight-person team went through Y Combinator's summer 2026 cohort.
The startup claims its platform has pushed call abandonment down to roughly 2% from the industry norm of 30%, and converts more than 80% of eligible patient calls into confirmed bookings. At Jobfit, Avoca resolved 86% of calls without human intervention. Those are impressive figures, if they hold across different practice sizes and patient populations.
Avoca's customer roster includes Australian radiology networks like Lumus Imaging, Western Radiology, ForHealth, and Precision Imaging Partners. The company says it now serves organizations operating over 150 locations. All patient data stays within Australian data centers under the current architecture, and Avoca lists SOC 2 Type II, ISO 27001, and HIPAA compliance on its security documentation.
The technical challenge is less about voice recognition than encoding an unruly tangle of clinical protocols and operational rules. Radiology scheduling is not like booking a restaurant table. A cardiac CT requires different preparation than a mammogram. Some scans need radiologist review of prior films before scheduling. Certain technologists have subspecialty training that must match the procedure type. Scanner availability shifts based on maintenance windows and emergency cases. Avoca's platform attempts to bake all of that institutional knowledge into software that can execute the logic in real time during a patient call.
Integrations span radiology information systems including Karisma, Comrad, Visage, Epic, Oracle Health's Cerner platform, and MEDITECH. The company says connections to athenahealth and eClinicalWorks are in the works. When something breaks or a case falls outside programmed parameters, the system escalates to staff via Microsoft Teams or Slack.
Avoca operates in a market where multiple players are chasing patient access automation with varying approaches. PocketHealth markets Conductor for radiology back-office workflows. Luma Health released Spark AI agents in recent years for outreach and booking, including fax parsing that generates EHR referrals. Notable Health focuses on closed-loop referral management with Epic integration. Syllable and Hyro offer conversational AI for scheduling across specialties. Aidoc has promoted an "AI OS" concept, though its focus leans toward clinical algorithm deployment in imaging rather than front-door patient logistics.
Other competitors like Rad AI, Radloop, and Within Health concentrate on the back end: tracking incidental findings and follow-up care rather than initial appointment booking.
The company has started pursuing American customers following its Y Combinator cohort. A July post from Shand requested introductions to radiology network executives and health system C-suite leaders. So far, Avoca has not disclosed U.S.-based customers or detailed how it will handle data residency requirements for protected health information under American HIPAA regulations, which demand business associate agreements and typically require hosting on domestic infrastructure.
The lack of specifics on U.S. deployment is understandable given the company's current focus on Australian operations. Breaking into the American healthcare market means navigating not just different compliance regimes but also fragmented electronic health record systems, varied payer rules, and entrenched vendor relationships. Australian success does not automatically translate across the Pacific, though the underlying problem Avoca targets exists in both markets.
Whether the startup can maintain its performance metrics as it scales beyond its home turf, and whether American patients prove as comfortable talking to its AI as Australian ones apparently are, remains an open question. For now, Avoca has carved out a niche solving a genuine operational headache in radiology practices. The real test comes when it tries to do the same thing at American scale.
