Australian startup Heidi Health has become one of the most visible examples of AI actually shifting the dial on healthcare productivity – and New Zealand is at the forefront of that story.
In this week’s episode of The Business of Tech, I talk to Heidi co‑founder Yu Liu about the company’s journey from student training tool to AI “care partner” for clinicians, and its audacious goal of doubling global healthcare capacity.
Heidi didn’t start life in the emergency department. Yu and his co‑founders first built Oscar, a chatbot that helped medical students practise exam skills – essentially simulated patients for training bedside manner and clinical questioning.
Oscar was useful, but the startup team struggled to find students willing to pay for it. In 2019, Liu and his co-founders, Dr Tom Kelly and Waleed Mussa, pivoted to tackling one of the biggest bottlenecks in healthcare – the hours clinicians lose every day to documentation and administration.
Widespread use in emergency departments
That created Heidi Scribe, an ambient AI scribe that sits in on consultations, listening to the conversation and producing high‑quality clinical notes tuned to each hospital’s templates and workflows. Clinicians were quick to adopt it. Liu describes doctors using Heidi in every consult and calling the founders directly when it went down, because they no longer wanted to go back to typing everything themselves.
In New Zealand, that enthusiasm has translated into national‑scale deployment. Health New Zealand is rolling Heidi out across all emergency departments, with clinicians in places like Hawke’s Bay cutting documentation time per patient from roughly 17 minutes to around four minutes.
Heidi, which has now raised around US$100 million across several VC-backed fundraising rounds, blends frontier large language models with specialised, region‑local models trained on clinical language and medication names, hosted in‑region to satisfy data sovereignty requirements. That’s how it pushes accuracy toward the near‑99 per cent threshold clinicians need to trust AI‑generated notes, says Liu. Heidi wants to transform assistive AI into something closer to infrastructure.
From scribe to evidence-gatherer
Heidi doesn’t retain recordings of conversations, and while many doctors create transcriptions on their smartphones or laptops, the Heidi Remote is also available – a mobile recorder doctors and nurses can carry around clinics and hospitals for easy recording that doesn’t rely on an internet connection.
The company is already moving beyond transcription. Heidi Evidence surfaces relevant clinical research and guidelines at the point of care, while Heidi is expanding into pre‑chart summaries, referrals and spoken commands that trigger real actions in electronic health record systems. The aim, says Liu, is to let doctors focus on diagnosis and human connection, and let AI handle everything else.
In the episode, we dig into Heidi’s founding story, its rapid uptake in New Zealand’s public health system, and the governance and privacy questions that come with putting AI in the consult room.
Tune in to the full conversation with Heidi Health co-founder Yu Liu on The Business of Tech, available wherever you get your podcasts.
See omnystudio.com/listener for privacy information.
New Zealand