Christchurch’s WeCare is running more than 30,000 enrolled patients on three GPs, one of them part-time. At the national average, which a 2025 analysis put at 74 GPs per 100,000 people, a practice that size would carry about 22 doctors.
That gap is the story. WeCare is not another clinic bolting a transcription tool onto a tired model. It is a live test of whether software can stretch the scarcest labour in the health system without turning the family doctor into a call centre.
The AI does the intake, a clinician signs off
The practice, run by brothers Dr Brian Ko and Dr John Ko, launched its AI assistant Quindlee in September. Built largely on ChatGPT and Claude, it takes a patient’s history and complaint, passes it to a back-office clinician, or routes it elsewhere if it cannot help. Clinicians approve prescriptions, care plans and medical certificates, sometimes without speaking to the patient. The scope covers around 40 common complaints, from acne, shingles and UTIs to weight loss managed with Wegovy and Mounjaro. Next year WeCare plans to push Quindlee into pre-consult history-taking and summarising.
This is a real step beyond what most of the sector has done. Back in 2024, RNZ reported that hundreds of GPs were turning to AI mainly to transcribe consultations and summarise notes. A 2025 primary care survey found 68% of staff were using AI tools, but scribing dominated at 50%, with chatbots and virtual assistants at just 16%. Everyone else was automating paperwork. WeCare is automating the front door.
The doctor count is falling as patients pile in
WeCare started in 2016 with 1,300 patients in one Canterbury clinic and, by mid-2025, had passed 23,000 across four branches. In August this year NZ Doctor reported 29,300 patients staffed by four GPs, two nurse practitioners and seven physician associates, with a fifth clinic planned.
Three months on, patient numbers are higher and the GP headcount is lower. To be fair to the conventional comparison, WeCare leans heavily on nurse practitioners and physician associates, so “three doctors” understates its clinical workforce. But that is precisely the point. The model pushes routine work down to the cheapest safe level of skill and keeps GPs for what only GPs can do.
It is also priced to scale. In January 2026 WeCare took its nurse-led CareNow service online nationwide at $25 for enrolled patients and $35 for casual ones, with Brian Ko saying at the time it “removes cost as a barrier while still being delivered by an experienced clinical team”.
Why this matters beyond the waiting room
For employers, a slow GP system is a productivity drain hiding in plain sight. Staff waiting a week for a medical certificate, a UTI script or a shingles diagnosis are staff off work or working badly. A $35 online consult with a fast turnaround is a straightforward business benefit, and insurers and HR teams should be watching whether it holds up.
The workforce maths makes the case for experimenting. The same 2025 analysis forecast the GP ratio would fall to 70 per 100,000 by 2031, against 116 in Australia. New Zealand will not train its way out of that gap quickly. Health NZ has set primary care access targets, but WeCare is already operating where the system says it wants to be.
Speed is easy, continuity is hard
The risk is not that Quindlee gets acne wrong. It is that care fragments. In 2025, Luke Bradford, then president of the Royal New Zealand College of General Practitioners, asked how technology could leave patients feeling supported “as opposed to transacting with a machine and depersonalising care, which we know actually leads to worse outcomes and splintered care”. He also warned against “creating a two-tier healthcare system or worse”.
Those are fair challenges, not Luddism. The 2025 survey flagged privacy, data integration and AI notes that were “overly verbose or inaccurate”. A chronic patient who never sees the same clinician twice is a patient whose deterioration can slip through.
The proof WeCare now owes the sector
WeCare has shown it can grow fast on thin GP labour. The next test is harder. It needs to publish outcomes, not just throughput: escalation rates, missed diagnoses, repeat presentations, patient continuity for chronic conditions. If those numbers stand up as Quindlee moves into history-taking in 2027 and CareNow Online spreads nationally, this becomes the template every practice owner and health funder copies. If they do not, it becomes the cautionary tale. Either way, the experiment is too important to judge on headcount alone.
Sources
- NZ Herald: WeCare doctors: The Christchurch brothers using AI to rethink the family practice (2026-11-07)
- NZ Doctor: Quick access leads to speedy growth for WeCare clinics (2026-08-11)
- Chris Lynch Media: Canterbury healthcare provider WeCare launches low cost online medical service nationwide (2026-01-06)
- RNZ: Hundreds of GPs turning to AI to help manage workloads (2024-07-04)
- GPNZ: Round two AI in Primary Care survey highlights growing adoption (2025)
- AI Researchers NZ: Forget robot doctors, what’s holding back AI in healthcare (2025-11-17)
Join the discussion
Add useful context, ask a good question, or challenge an idea — keep it specific and respectful.