The math was brutal, and radiology executives in Australia knew it. Three out of every ten patients who called to schedule an MRI would give up before anyone picked up. Hold times stretched, phone lines jammed, and somewhere in that mess of missed connections were delayed diagnoses, lost revenue, and a patient experience that felt, well, broken.
Enter Avoca Systems, an eight-person startup out of San Francisco with a specific thesis: radiology is different. Different enough that the usual patient access software won't cut it. Different enough to warrant an AI platform built specifically for the operational chaos of imaging networks.
Founded by Max Shand and Campbell Mercer-Butcher, Avoca emerged from Y Combinator. The company says it now serves more than 250 clinics across several of Australia's largest radiology networks, reporting call abandonment reductions from around 30% down to 2%—figures the company has not independently verified. Patient satisfaction scores, according to Avoca's own data, sit at 4.6 out of 5.
Those are the kinds of numbers that get attention. Now, fresh from demo day, the startup is turning its sights toward the U.S. market—a vastly more complex landscape, but one wrestling with the same fundamental problem.
More Than Just Scheduling
What Avoca has built isn't appointment software in the conventional sense. It's an automation layer—an "operational intelligence" platform, as the company describes it—that handles the entire patient intake workflow for radiology clinics.
A voice AI agent fields inbound and outbound calls. The system processes online bookings, ingests electronic referrals and faxes (yes, healthcare still runs on fax), and manages ongoing patient communications. All of this sits on top of existing radiology information systems and electronic health records, writing appointment data back in real time.
Here's where it gets complicated, and why Shand and Mercer-Butcher argue generic solutions fall short: scheduling a cardiac MRI isn't like booking a routine checkup. The platform has to encode clinic-specific rules, clinical protocols, staff rosters, site constraints, equipment availability. When a patient calls, the system checks referral requirements, asks clinical safety questions, matches the request to radiologists with the right subspecialty training and technologists certified for that particular procedure, verifies scanner availability, then books the appointment directly into the RIS.
Avoca reports booking conversion above 80% on eligible calls, though those figures—like much of Avoca's performance data—come from the company itself rather than independent validation.
The Australian Laboratory
According to Avoca, its customer roster in Australia includes major radiology networks: Lumus Imaging, Partnered Health, Western Radiology, ForHealth, Precision Imaging Partners, Ochre Health. The company points to case study metrics from Partnered Health's Jobfit brand showing an 86% call resolution rate alongside those satisfaction scores.
But perhaps the more telling detail is the integration footprint. Avoca works with the RIS systems common in Australian radiology—Karisma from Kestral, Aura from Comrad, Visage from Pro Medicus. On the EHR side: Epic, Oracle Health (formerly Cerner), MEDITECH. Integrations with athenahealth and eClinicalWorks are listed as coming soon. The platform also hooks into Microsoft Teams, Slack, and handles telephony over SIP and PSTN.
The company maintains SOC 2 Type II, ISO 27001, and HIPAA compliance certifications. Patient data currently stays in Australia under the platform's data residency architecture—a structure that will presumably need rethinking as Avoca expands into the U.S.
The Afterpay Connection
Shand's background offers some context for the ambition here. He was the first employee at Afterpay, the buy-now-pay-later fintech that became one of Australia's most celebrated startup stories before Block (formerly Square) acquired it for $29 billion in 2022. After that, he founded and exited Serenade, a music merchandise and NFT platform.
Mercer-Butcher brings technical depth from his time as Principal Software Engineer at Honey Insurance and Tech Lead at Ailo. He also previously founded a company called Epex.
The pair bet on a vertical approach: solve one industry's unique complexity really well, rather than building horizontal patient access software that tries to serve every specialty. An orthopedic clinic schedules based on provider availability and maybe some prep requirements. Radiology networks juggle exam types, specific equipment, radiologist subspecialties, technologist certifications, contrast protocols, patient safety screening, insurance authorization, and referral requirements that shift depending on payer and exam type. Then layer in coordination across multiple sites with different hours and capabilities.
It's a headache, operationally speaking. And headaches, historically, are what startups try to solve.
Crossing the Pacific
The U.S. market presents a different beast entirely. Healthcare fragmentation means more integration points, more payers, vastly more prior authorization complexity. But it also means larger networks with bigger operational pain points.
Policy momentum might be working in Avoca's favor. CMS has been pushing for electronic prior authorization, aiming to reduce reliance on fax and portal-based workflows ahead of payer requirements set to take effect. The shift toward standardized digital referral intake aligns with Avoca's positioning as an automation layer for e-referrals and fax processing.
The competitive landscape isn't empty. Established players like AbbaDox have announced operational AI features. RamSoft offers AI scheduling modules. Broader patient engagement platforms like Phreesia and Relatient operate in adjacent territory. Newer entrants such as Scope Health and Carethink are building similar patient access automation tools, though most take a horizontal approach across specialties rather than Avoca's radiology-specific bet.
The timing is notable, if nothing else. A broader wave of operational AI platforms is washing into radiology. Sirona Medical and Everlight Radiology announced an "AI-native" RadOS partnership. Aidoc received FDA Breakthrough Device designation for AI that drafts chest X-ray reports. The pattern suggests radiology networks are at least receptive to AI-powered workflow automation—provided it integrates cleanly with existing systems and delivers measurable ROI.
Whether Avoca's Australian playbook translates stateside remains an open question. The company hasn't yet announced signed U.S. customers or pilots. But the fundamentals are familiar enough: too many phone calls, not enough staff, scheduling complexity that overwhelms manual processes, and relentless pressure to improve patient access without ballooning costs.
If the numbers from Australia hold in the States—and that's a meaningful if—radiology network operators will be watching closely. Because 30% call abandonment isn't just a metric. It's patients who gave up, scans that didn't happen, and a system straining at the seams. Sometimes, the most valuable innovation isn't the flashiest. It's the one that picks up the phone.
