In offices, hospitals, and clinics across 116 countries, an Australian health-tech startup is quietly transforming the way clinicians document patient visits. Heidi Health, co-founded by vascular surgeon Dr Thomas Kelly, is using artificial intelligence to shoulder the administrative load — freeing up doctors to care, not type. Its recent fundraise, global adoption, and expansion into veterinary care mark a pivotal moment not just for Heidi, but for health systems struggling under rising demand.
The Administrative Avalanche
Doctors spend an alarming share of their time on paperwork. Some studies suggest clinicians devote equal or more hours to documentation and bureaucracy than to seeing patients. This imbalance contributes to burnout and threatens system capacity. Heidi Health steps into that breach by automating key tasks: transcribing conversations, generating clinical notes, crafting referral letters, summarising previous visits, and drafting follow-up plans. (Heidi Health)
Dr Kelly’s drive to build Heidi came from first-hand frustration. In his surgical and outpatient stints, he witnessed how clinicians stayed late — “pyjama time”— to catch up on documentation. The strain was not sustainable in a system under pressure from ageing populations, rising chronic disease, and a looming shortage of general practitioners. (SmartCompany)
“Building a sustainable healthcare system requires expanding clinical capacity without compromising clinician wellbeing or patient safety,” he has said. (The Australian) Heidi’s mission codified: double global healthcare capacity by liberating clinicians from paperwork. (Heidi Health)
Funding, Scale, and Market Momentum
In 2025, Heidi closed a US$65 million (AUD 98 million) Series B round, led by Point72 Private Investments, with backing from Blackbird Ventures, Headline, and LocalGlobe. (SmartCompany) That capital lift bumped its valuation to around US$465 million (some sources report AUD 704 million) — a stunning rise from its earlier valuation of US$100 million just months prior. (SmartCompany) Heidi now handles over 2 million consultations weekly and has processed more than 73 million consults in its lifetime, in 110 languages. (The Australian) The platform’s network effect is real: clinicians recommend it to peers; institutions adopt it broadly. (SmartCompany)
In earlier stages, Heidi had raised ~$10 million AUD in its Series A, led by Blackbird Ventures and associated investors. (SmartCompany) A top-up of ~AUD 26 million (USD 16.6 million) followed in 2025, led by Headline. (SmartCompany) Those funds went toward expanding the team, refining models, and accelerating global expansion. (SmartCompany)
Heidi’s pace is now compared to Canva’s early growth — one of the fastest in Blackbird’s portfolio. (Capital Brief) Its ability to scale quickly, despite regulatory, privacy, and clinical complexity, sets it apart. (Capital Brief)
How Heidi Works: Core Technology & Workflow
At its heart, Heidi is an AI medical scribe. During a patient consultation, Heidi captures audio (with patient and clinician consent), transcribes it, and then applies algorithms and healthcare models to generate structured clinical notes and supporting documents. (Heidi Health) It supports templates and customization — users can tailor note style, macros, or prompts to match their preferences. (Heidi Health) In many cases, clinicians correct or tweak Heidi’s first draft, which then “learns” and improves over time. (dayone.fm)
Heidi also experiments with pre-consult modules: patients can input their symptoms or history in advance. The platform then structures it for clinician review before the visit, reducing time spent on intake. (SmartCompany) Over time, Heidi hopes to assist in prompting clinicians with reminders or alerts — for example, flagging missed actions from prior visits. (SmartCompany)
To ensure scalability and independence, Heidi is developing its own machine learning models, moving away from third-party providers. Currently it uses Microsoft and OpenAI tools as part of the stack, but anticipates full in-house processing. (SmartCompany) Data privacy is central: patient data is de-identified, double encrypted, and stored on local servers when possible. Heidi’s architecture is built to comply with clinical, privacy, and regulatory requirements. (SmartCompany)
Dr Kelly has been cautious about overstepping clinical boundaries: Heidi currently does not generate treatment plans or make diagnoses, though he hasn’t ruled out extensions in the longer term. (ausdoc.com.au) He frames Heidi as a “partner” to clinicians, not a replacement. (ausdoc.com.au)
Real-World Adoption & Case Studies
Australia
In its home market, Heidi is already used in major public systems. Monash Health and Queensland Health have integrated the tool in facilities across their networks. (SmartCompany) Heidi’s local presence anchors its legitimacy in a complex regulatory environment.
Global Systems
Heidi has secured partnerships with health systems abroad: NHS trusts in the UK, Beth Israel Lahey Health in the US, and the Yukon government in Canada are among its collaborators. (The Australian) These deals facilitate deeper integration into care delivery systems beyond solo clinics. (The Australian)
Veterinary Expansion
In a surprising yet strategic move, Heidi has branched into veterinary healthcare via a partnership with Greencross, Australia’s largest veterinary network. (The Australian) About 35% of Greencross consultations now use Heidi — some clinics even exceed 60% usage. (The Australian) The reason? Vets often have their hands full and cannot type mid-examination; automating notes makes immediate sense in that context. (The Australian) Dr Kelly also sees it as a trust bridge: “people might trust AI for pet care first, then for themselves.” (The Australian)
Risks, Challenges & Ethical Dimensions
Regulation is inevitable in healthcare, and AI in medicine is under scrutiny. In some jurisdictions, AI tools face barriers related to liability, auditing, and clinical safety. Heidi must navigate a patchwork of regulation in each market it enters. In addition, algorithmic bias, data security, patient consent, and transparency are vital guardrails. Dr Kelly acknowledges these trade-offs and insists on clinician oversight, not autonomous decision-making. (ausdoc.com.au)
Clinicians may resist too much automation in clinical thinking — fearing “black boxes” or erosion of professional skill. Heidi’s approach to transparency, customization, and clinician review intends to mitigate that. (dayone.fm) Competition is intensifying. Microsoft has introduced Dragon Copilot aimed at doctors, which may encroach on Heidi’s territory. (The Australian) Heidi must maintain differential edges in trust, safety, and domain knowledge.
Another open question: revenue model. Heidi aims to serve solo clinicians (who choose it) and enterprise clients licensing the platform. But its long-term viability depends on sustainable pricing, retention, and defensibility. (mobihealthnews.com) A further technical challenge is the transition from reliance on third-party LLMs to stable, robust in-house models, with lower costs and better controllability. (SmartCompany)
Concrete Steps for Clinicians and Health Leaders
If you are a clinician curious about Heidi (or a similar AI scribe), here’s a roadmap:
- Request a pilot or trial: Many AI health techs offer optional trial periods. Use this to compare existing note workflows vs. AI-augmented output.
- Evaluate templates and customization: The more flexibility a tool offers to mirror your style, the more adoption.
- Involve IT & privacy teams early: Confirm encryption, data residency, de-identification protocols, and compliance with local health regulations.
- Monitor clinician feedback closely: Track time savings, note quality, and error rates. Adjust human editing workflows dynamically.
- Plan change management: Introduce the tool gradually. Train champions among staff, gather feedback, iterate.
- Benchmark metrics: Time saved per consult, reduction in after-hours documentation, clinician satisfaction, and patient impact.
- Guard against overreliance: Always maintain clinician oversight and audit random outputs for safety.
- Consider integration: Over time, linking the tool to EMRs or EHRs further streamlines workflows — but beware integration complexity.
For organizational decision makers:
- Conduct ROI modelling: Compare clinician time saved vs. licensing costs.
- Start with high-burden specialties: Areas like internal medicine, chronic disease, pediatrics (lots of history) yield larger gains.
- Pilot in non-critical units: Outpatient clinics, allied health, or emergency documentation are lower-risk test beds.
- Use the data stream: Aggregated, anonymised insights from multiple clinicians can help spot patterns, gaps, quality issues (if privacy is managed).
- Negotiate rights to customization and data export: Ensure you’re not locked into closed systems.
Why This Matters for Health Systems
Healthcare supply can’t keep up with growing demand. Admin burdens throttle effective output. The global shortage of clinicians means every minute matters. Heidi’s ambition to reclaim clinician hours could shift health economics.
By taking repetitive tasks off human shoulders, tools like Heidi allow doctors to focus on diagnosis, patient connection, and decision-making. If deployed broadly, the cumulative effect could be a measurable uplift in system capacity and patient access.
Perhaps most importantly, Heidi acts as a bridge technology: it doesn’t try to replace the clinician’s judgment but augments it. That model is more likely to gain trust and acceptance.
Heidi’s veterinary expansion is also strategic: it builds user trust in a lower-stakes domain, cross-pollinating acceptance into human care. Its funding surge proves investor confidence in its model and pace.
The short term will test its product robustness, clinician sentiment, regulation, and competition. But if it navigates those hurdles well, Heidi could be a case study in how Australia built global health impact from local roots.
In the coming years, Heidi — and platforms like it — may redefine how clinical care is documented, delivered, and scaled. Expect more clinicians to ask: can I offload that note? And more systems to fund it.
