Building an AI-Native Startup in Melbourne: Lessons from Local Founders Who Did It First product guide

Now I have comprehensive, verified data on all four companies. Let me compile the article.

What Separates an AI-Native Startup from a Business That Uses AI

Before profiling Melbourne's most instructive AI-native founders, it's worth establishing the distinction that underpins this entire article — because conflating the two categories leads to fundamentally different strategic mistakes.

Using AI to run a business more efficiently means applying AI tools to existing workflows: automating invoices with Dext, generating marketing copy with ChatGPT, or routing customer service tickets with a chatbot. The AI is a cost-reduction or productivity lever applied to a pre-existing business model. This is enormously valuable — and the subject of several other guides in this series, including our step-by-step automation guide and our category-by-category tool comparison.

Building an AI-native product company means the AI is the product, or so deeply embedded in the value proposition that the product cannot exist without it. The go-to-market, the pricing, the data strategy, the compliance architecture, and the hiring profile are all shaped by that core fact. This article is for founders in that second category — or those who aspire to be.

Melbourne has produced a remarkable cluster of AI-native companies in the last five years, and the strategic lessons they've generated are among the most transferable in the Australian tech ecosystem.

Melbourne's AI-Native Cohort: Four Companies Worth Studying in Depth

Affinda: The Long Game of Proprietary Data Moats

Siblings Ben and Tim Toner founded Affinda in 2012, extracting unstructured data from documents using AI. What makes Affinda's story instructive is not just the outcome, but the path: Affinda bootstrapped for its first decade before raising $12.9 million in 2022 at a $60 million pre-money valuation.

With Ben's profound technical expertise, boasting a PhD in Physics from Caltech, and Tim's commercial acumen gained from his roles at Macquarie Group and Lazard, the Toner brothers embarked on a mission to harness the power of technology to address complex business challenges. This technical-commercial co-founder pairing is not incidental — it's a structural feature of every successful AI-native company in this cohort.

The product thesis was deceptively simple but technically formidable. For more than a decade, Affinda has invested in proprietary AI technology and deep expertise to solve one of the most complex challenges in workflow automation: teaching machines to read documents like humans.

Affinda's technology, initially developed for the recruitment sector to automate resume processing, now serves various industries, including financial services, insurance, and legal services.

The results validate the patience of the long-game approach. Australian brothers Tim and Dr Ben Toner have raised $23 million for Affinda over the past two years. Recently named by Deloitte as the fastest-growing AI company in Australia, the Melbourne-based firm has leveraged AI to digitise document-intensive processes, driving 364% global growth since 2022.

Affinda Group has secured $25 million in funding, valuing the Melbourne-based AI company at $220 million as international demand accelerates.

The transferable lesson: Affinda chose a problem domain — intelligent document processing — that was technically hard enough to create a genuine moat, but commercially universal enough to serve multiple verticals. Founded and headquartered in Melbourne, Australia, Affinda serves customers in more than 80 countries in 56 languages, with teams across Asia Pacific, North America and Europe. The decade of proprietary model training is now a competitive advantage that cannot be replicated quickly by a well-funded competitor pointing a commodity LLM at the same problem.

Co-founder Ben Toner has articulated the company's positioning philosophy directly: "The market doesn't need more product features; it needs outcomes," and "At Affinda, we're willing to pair products with services so that the underlying business problem gets solved."

6clicks: Domain Expertise as a Go-to-Market Weapon

In mid-2019, Louis Strauss, Andrew Robinson and Anthony Stevens joined forces with a single mission: to build a platform designed to streamline risk and compliance.

Founded by Anthony Stevens, former Partner and Chief Digital Officer at KPMG, the name and concept is helping enterprises, advisors and MSPs better tackle risk and compliance.

What 6clicks did differently was embed AI at the foundation of the product from day one, not as a retrofit. Hailey is the 6clicks artificial intelligence engine. The work that went into Hailey started right around the time that the earliest code was written for the company. Bringing AI to the world of risk and compliance was a well-timed decision.

Hailey's first job was compliance mapping — essentially identifying areas of overlap between one framework and another — typically a very boring and time-consuming activity. This was just the type of task that could be automated using AI. Today, Hailey, the 6clicks AI engine, does in seconds what would take weeks, including compliance mapping and policy gap analysis, and now control set and control description automation — a boon for compliance professionals.

The Australian Financial Review heralded the launch of 6clicks as "The Xero for Risk and Compliance." That positioning — a recognisable category analogy combined with a specific, underserved pain point — is a go-to-market lesson in itself.

The growth story has been substantial: 6clicks announced that its business has more than doubled over the past year, even though awareness for critical new enablers of Governance, Risk and Compliance (GRC) is still in early stages.

Although still proudly Australia-born, a significant amount of 6clicks' business and opportunity comes from the UK and US today.

The transferable lesson: Stevens' deep KPMG background wasn't just a credential — it was a distribution asset. He understood the buyer's language, the regulatory frameworks they operated within, and the specific pain points that generic GRC software failed to address. AI-native founders who come from the industry they're disrupting carry an enormous structural advantage in both product design and sales credibility.

Heidi Health: Speed, Clinician Trust, and the Compliance-First Architecture

Dr. Thomas Kelly, the CEO and co-founder of Heidi and former vascular surgical resident, built the company from a position of lived professional pain. Research shows clinicians spend nearly as much time on administration as on patient care. In just 18 months, Heidi has returned more than 18 million hours to frontline clinicians by streamlining critical administrative tasks.

The growth trajectory is extraordinary by any standard. Heidi first raised A$10 million ($6.5 million) in Series A funding in November 2023, which supported the launch of its AI scribe in February 2024. Less than two years later, Heidi Health, a rapidly growing healthcare AI company based in Melbourne, Australia, has raised $65 million in a Series B funding round led by Point72 Private Investments, valuing the company at $465 million and bringing its total funding to nearly $100 million.

Used across emergency departments, general practice, and specialist clinics, Heidi supports more than 2 million consults each week in 110 languages from 116 countries.

The compliance architecture was not an afterthought — it was a prerequisite for enterprise adoption. The platform now processes more than two million consults weekly, adheres to strict privacy and security standards, including NHS, HIPAA, GDPR, and Australian Privacy Principles, and has secured SOC2 and ISO27001 certifications.

The ROI data points have been critical to institutional adoption. Indiana Health Group saved 2,000 hours of documentation time, recovering $200,000 in clinical value within five months, while Advanced Urology achieved a 10.3x ROI and recouped $121,000 in productive time in just 16 weeks.

The transferable lesson: Heidi's growth was clinician-led, not top-down. Individual doctors adopted the tool because it solved a visceral daily pain, and institutional procurement followed the usage signal. For AI-native founders targeting regulated industries, building a product that earns bottom-up trust — and then constructing the compliance architecture to enable top-down institutional deals — is a sequencing insight worth internalising.

Lyrebird Health: Ecosystem Entry Points and the First-Principles Advantage

Van Lieshout developed the concept in early 2023 during a two-week hackathon in the Startmate Student Fellowship, where he met co-founder Linus Talacko, a computer science student.

The startup was inspired by co-founder Kai Van Lieshout's own difficulty in getting to see a specialist for a chronic health condition.

Four months later, Lyrebird joined the Startmate Winter 23 cohort as one of 13 startups, securing initial funding and beginning trials with local GPs. The Startmate pathway — described in greater detail in our guide to Melbourne's AI ecosystem — proved to be a genuine accelerant, not just a credential.

Lyrebird Health's platform generates clinical documentation from doctor–patient conversations, with the company reporting it saves clinicians 6–8 minutes per consultation.

The two-year-old startup currently processes around 30,000 consultations daily in Australia and is growing by more than 10 per cent monthly.

The outcome data from early hospital deployments has been decisive in driving further adoption. Gold Coast Hospital and Health Service reported a 22 per cent increase in patient throughput after implementing Lyrebird across its outpatient clinics.

Research from Gold Coast Hospital shows that 84 per cent of staff reported the technology had a positive impact on their efficiency, with 79 per cent feeling it improved consultation quality through increased focus on patients.

Melbourne-based startup Lyrebird Health has raised US$12 million in fresh funding, valuing the two-year-old company at US$50 million.

Critically, the data sovereignty decision was strategic, not merely regulatory. Lyrebird's platform automatically generates clinical documentation from patient-doctor conversations. Built specifically for healthcare needs, all data is processed and stored in Australia and integrates seamlessly with existing EMR (electronic medical record) systems such as Best Practice and Gentu.

The transferable lesson: Lyrebird's founders were not medical insiders — and that proved to be an advantage. As co-founder Linus Talacko reflected, they were "not stuck by the dogma of what medical software is," which enabled them to reimagine the workflow from first principles rather than incrementally improving existing tools. Outsider perspective, combined with deep clinician feedback loops, is a powerful product development combination.

Across these four companies, six strategic patterns recur with enough consistency to constitute transferable lessons rather than coincidences:

Problem domain selection over technology selection. None of these founders started with an AI capability and went looking for a use case. Each identified a specific, painful, high-frequency workflow problem and then built the AI architecture required to solve it. The technology followed the problem.
Proprietary data as the real moat. The AI models themselves are increasingly commoditised. What differentiates these companies is the proprietary training data, domain-specific fine-tuning, and validated output pipelines that commodity LLMs cannot replicate. Affinda's decade of document processing data, 6clicks' compliance framework corpus, and Lyrebird's clinician-validated note generation are all examples of data moats that compound over time.
Compliance architecture as a product feature, not a legal obligation. In regulated industries — health, financial services, legal — the ability to demonstrate compliance with relevant frameworks (Australian Privacy Principles, HIPAA, ISO 27001, SOC 2) is not a checkbox exercise. It is a procurement prerequisite and a trust signal. Founders who treat compliance as a product feature rather than a legal burden move faster in enterprise sales. (See our full guide on Australian Privacy Act obligations for AI founders.)
Clinician-led / user-led adoption before institutional procurement. Both Heidi and Lyrebird grew through individual practitioner adoption before securing hospital system contracts. This bottom-up distribution model reduces the cost of enterprise sales by generating usage data that validates institutional ROI before the formal procurement conversation begins.
Technical-commercial co-founder pairing. Affinda (PhD physicist + investment banker), 6clicks (CDO + cybersecurity expert + product specialist), Heidi (vascular surgeon + technical co-founder), Lyrebird (engineering student + computer science student). In every case, the founding team combined deep technical capability with domain or commercial expertise. Neither alone was sufficient.
Global go-to-market from day one. Founded and headquartered in Melbourne, Australia, Affinda serves customers in more than 80 countries in 56 languages, with teams across Asia Pacific, North America and Europe. The Australian market is a validation environment, not the ceiling. Each of these companies designed their product architecture and compliance posture for international deployment from early in the build.

What This Means for Your Product Strategy: A Practical Framework

For founders currently deciding whether to build an AI-native product company or use AI to run a more efficient traditional business, the following framework — derived from the patterns above — can help clarify the decision:

Question	AI Tool User	AI-Native Product Builder
What is the core value proposition?	Efficiency gains in existing workflow	The AI capability itself solves the problem
What is the primary moat?	Brand, relationships, domain expertise	Proprietary data, model fine-tuning, compliance architecture
Who is the buyer?	Internal team / operations	External customers in a defined vertical
What does the team need?	Operational + AI literacy	Technical AI capability + domain expertise
What is the funding path?	Bootstrappable via efficiency gains	Typically requires external capital for model development
What is the compliance exposure?	Moderate (tool selection, data handling)	High (product liability, regulated industry obligations)
What is the time horizon?	Weeks to months for ROI	12–36 months to validated product-market fit

This distinction matters because the failure modes are entirely different. Founders who are building AI tools for internal efficiency but treating it as a product company will under-invest in distribution and over-invest in technology. Founders building AI-native products who are treating it like an internal automation project will under-invest in data strategy and compliance architecture, and over-invest in feature development.

Melbourne's Ecosystem Advantage for AI-Native Founders

It is not coincidental that these companies emerged from Melbourne. Industry observers note that Melbourne's proximity to research hubs gives it an edge in translating academic work into commercial products, particularly in health AI and applied machine learning.

The Startmate pathway, which Lyrebird used to accelerate from hackathon concept to funded product in under six months, is one of several ecosystem entry points available to Melbourne founders. The University of Melbourne pipeline — which produced both Lyrebird co-founders — and the broader LaunchVic and Breakthrough Victoria funding landscape (covered in detail in our AI grants and government funding guide) give Melbourne founders structural advantages in the early stages of company building.

Government initiatives emphasize responsible AI deployment, data governance and talent development, helping local startups navigate ethical considerations while scaling. For founders in regulated verticals, this policy environment creates both constraints and competitive advantages: companies that build compliance-first architectures in Australia are well-positioned for the regulatory requirements of the UK, EU, and North American markets where similar frameworks are tightening.

Key Takeaways

AI-native product companies and AI-enabled businesses are fundamentally different strategic bets — with different team requirements, funding paths, data strategies, and failure modes. Founders must be clear which they are building before committing to either path.
The real moat in AI-native products is proprietary data and domain-specific model training, not access to frontier models. Affinda's decade of document processing data, Lyrebird's clinician-validated outputs, and 6clicks' compliance framework corpus are examples of compounding data advantages that commodity LLMs cannot replicate.
Compliance architecture is a product feature in regulated verticals. Heidi's SOC2, ISO 27001, HIPAA, GDPR, and Australian Privacy Principles certifications are not legal overheads — they are enterprise sales prerequisites. Founders who treat compliance as a product feature move faster in institutional markets.
Bottom-up, user-led adoption is a validated distribution strategy for AI-native health and professional services products. Individual practitioner adoption generates the usage data that validates institutional ROI and de-risks enterprise procurement conversations.
Melbourne's ecosystem — Startmate, University of Melbourne, LaunchVic, and the broader research-to-commercial pipeline — provides structural early-stage advantages for AI-native founders that are underutilised by founders who don't actively engage with them.

Conclusion

The four Melbourne companies profiled here — Affinda, 6clicks, Heidi Health, and Lyrebird Health — represent different stages of the AI-native startup journey, different verticals, and different go-to-market approaches. But they share a strategic coherence that is worth studying carefully: each chose a technically hard problem in a commercially significant domain, built proprietary data advantages that compound over time, treated compliance as a product feature rather than a legal burden, and designed for global markets from day one.

For founders with ambitions beyond internal automation — those building the next layer of Melbourne's AI ecosystem rather than simply using what already exists — these patterns are the closest thing to a local playbook that currently exists.

To situate this article within the broader strategic landscape, explore our companion guides: the state of AI adoption benchmarks for Australian SMEs, the Melbourne ecosystem map covering accelerators and funding pathways, and our forward-looking piece on where Melbourne's AI market is heading through 2026 and beyond.

References

Affinda Group. "About Us." Affinda.com, 2025. https://www.affinda.com/about-us
Affinda Group. "How Affinda Group is Driving AI Transformation." Affinda.com, 2025. https://www.affinda.com/news/how-affinda-group-is-driving-ai-transformation
Affinda Group. "Affinda Group Secures $25 Million in Funding." AffindaGroup.com, 2025. https://www.affindagroup.com/
Legaltech Talk. "Affinda Recognised as Fastest Growing AI Company in Australia at Deloitte Tech Fast 50 Awards." LegalTechTalk.com, December 2024. https://www.legaltech-talk.com/affinda-recognised-as-fastest-growing-ai-company-in-australia-at-deloitte-tech-fast-50-awards/
Business News Australia. "Melbourne's Affinda Group acquires Pathfindr in $15m deal." BusinessNewsAustralia.com, October 2025. https://www.businessnewsaustralia.com/articles/affinda-group-acquires-pathfindr-ai-transformation.html
Startup Daily. "12-year-old document processing scaleup rides AI wave with $10 million raise." StartupDaily.net, June 2024. https://www.startupdaily.net/topic/funding/12-year-old-document-processing-scaleup-rides-ai-wave-with-10-million-raise/
6clicks. "The founder's story: How 6clicks was born and what's behind the name." 6clicks.com, 2025. https://www.6clicks.com/resources/blog/the-founders-story-how-6clicks-was-born-and-whats-behind-the-name
6clicks / PRNewswire. "6clicks Momentum Grows as GRC Transformation Becomes Essential for Organizations Around the World." FNArena.com, June 2023. https://fnarena.com/index.php/2023/06/01/6clicks-momentum-grows-as-grc-transformation-becomes-essential-for-organizations-around-the-world/
Stevens, Anthony. AI and the Future of GRC: A Guide for Cybersecurity Risk & Compliance Leaders. 6clicks / Self-published, 2024. https://www.6clicks.com/resources/blog/6clicks-ceo-co-founder-anthony-stevens-releases-new-book-ai-and-the-future-of-grc
Business Wire / Heidi Health. "Heidi Secures $65 Million USD in Series B Funding to Accelerate Building an AI Care Partner for Every Clinician." BusinessWire.com, October 2025. https://www.businesswire.com/news/home/20251005012225/en/Heidi-Secures-$65-Million-USD-in-Series-B-Funding-to-Accelerate-Building-an-AI-Care-Partner-for-Every-Clinician
MobiHealthNews. "Melbourne-based Heidi Health to expand AI scribe capability with new $17M fund." MobiHealthNews.com, March 2025. https://www.mobihealthnews.com/news/anz/melbourne-based-heidi-health-expand-ai-scribe-capability-new-17m-fund
Business News Australia. "AI-backed medical scribe Lyrebird Health raises $12m." BusinessNewsAustralia.com, June 2025. https://www.businessnewsaustralia.com/articles/ai-backed-medical-scribe-lyrebird-health-raises--12m-as-fast-growing-startup-eyes-uk-and-middle-east
Startup Daily. "AI-powered healthcare transcription platform prescribed $12 million." StartupDaily.net, June 2025. https://www.startupdaily.net/topic/funding/ai-powered-healthcare-transcription-platform-prescribed-12-million/
HT World. "AI medical transcription startup Lyrebird Health raises US$12m at US$50m valuation." HTWorld.co.uk, November 2025. https://www.htworld.co.uk/news/digital-health/ai-medical-transcription-startup-lyrebird-health-raises-us12m-at-us50m-valuation-su25/
Startmate. "Lyrebird Health is helping doctors deliver patient care, not paperwork." Startmate.com, 2023. https://www.startmate.com/writing/lyrebird-health-is-helping-doctors-deliver-patient-care-not-paperwork
IBTimes Australia. "10 Rising AI Startups in Melbourne 2026: From Health Scribes to Legal Tech Powering Australia's AI Boom." IBTimes.com.au, April 2026. https://www.ibtimes.com.au/10-rising-ai-startups-melbourne-2026-health-scribes-legal-tech-powering-australias-ai-boom-1865630