Two decades after Luke Anear swapped his career as a private investigator for a software idea sketched out in regional Queensland, the company he built is squaring up to one of the biggest tests of its existence. SafetyCulture, the workplace operations and safety platform that began life in Townsville in 2004, is now betting that the wave of artificial intelligence washing over enterprise software will lift it rather than drown it.
In an interview with The Australian Financial Review, Luke Anear laid out a case for why the business is geared up for an AI-first world, describing ambitions that stretch from underwriting insurance claims to running customer service through autonomous software agents. It is a notably confident posture at a moment when plenty of founders are hedging, and the AFR framed him as defiant in the face of an industry that is anxious about being hollowed out by the very technology it is racing to adopt.
From garage idea to global platform
SafetyCulture is best known for the mobile inspection and checklist tools that millions of frontline workers use to log hazards, complete audits and flag problems from factory floors, building sites, kitchens and warehouses. The pitch has always been that the people doing the work hold the most valuable data about how an operation actually runs, and that capturing it cleanly makes a business safer and more efficient. Over the years the company has layered on sensors, training, issue management and analytics, turning a single app into a broader operations platform used across the globe.
That installed base is the foundation for the AI ambitions Anear is now describing. A platform that already sits in the hands of frontline teams, collecting structured information about what is happening on the ground, is in theory well placed to feed AI systems that can summarise, predict and eventually act. The leap he is signalling is from software that records and reports to software that does the work, the shift the industry has taken to calling agentic AI.
The agentic bet, and the insurance angle
The two areas Anear singled out are telling. Agentic customer service points to a future where AI agents handle routine queries and tasks end to end rather than simply drafting a suggested reply for a human to approve. For a company serving large numbers of customers across many time zones, automating that layer promises both lower costs and faster service, provided the agents behave reliably enough to be trusted without constant supervision.
The move into underwriting insurance claims is the more ambitious play. SafetyCulture already holds a rich picture of how its customers manage risk, and the logic is that an operator who runs disciplined inspections and closes out hazards quickly is a better insurance bet than one who does not. Turning that safety data into pricing, cover or claims decisions would push the company from selling software into selling financial products, a very different business with its own regulatory weight. It is the kind of expansion that can dramatically enlarge a company’s addressable market, and also the kind that invites hard questions about data use, fairness and accountability.
Confidence versus caution
Anear’s optimism runs against a more cautious current in the software world. The prevailing worry is that general-purpose AI models from the largest labs could erode the value of specialised software by making it trivial for customers to build their own tools, or by letting a handful of dominant assistants absorb tasks that once required a dedicated product. Under that view, vertical software companies risk becoming a thin layer on top of commoditised intelligence, squeezed on price and stripped of their defensibility.
The counter-argument, and the one Anear appears to be leaning on, is that the moat is the data and the distribution, not the model. Foundation models are increasingly available to everyone, so the advantage shifts to whoever owns the proprietary information and the relationships with the people who generate it. On that reading, a company with years of frontline operational data and millions of users is holding exactly the assets that make AI useful in the real world, rather than impressive only in a demo. Which of these views proves right will not be settled by rhetoric, but by whether customers see measurably better outcomes from the agents being deployed.
What it means for Australia
SafetyCulture matters to the local technology story well beyond its own balance sheet. It is one of a small cluster of Australian software companies that have scaled internationally from unglamorous beginnings outside the Sydney and Melbourne hubs, and it has long been held up as proof that globally competitive software can be built from regional Australia. How it navigates the AI transition will be read as a signal for a crop of home-grown enterprise players now wrestling with the same question: whether AI is an accelerant for established platforms or a threat that favours whoever moves fastest and spends biggest.
The insurance ambition also lands in a sensitive policy environment. Australian regulators and the public have grown markedly more alert to how AI makes decisions that affect people and businesses, from credit to healthcare to employment. A platform using operational data to underwrite risk would sit squarely inside debates about transparency, consent and automated decision-making that governments here are still working through. For a company that trades on trust with frontline workers, getting the governance right is not a side issue, it is central to whether the strategy flies.
What’s next
The near-term test is execution. Talking about agentic service and AI-driven underwriting is one thing; shipping products that customers rely on, and that behave safely at scale, is another. Investors and rivals will be watching for concrete releases, early customers and evidence that the AI features drive real retention and spending rather than headlines. For a founder who has spent more than twenty years proving that an Australian software company can win on the world stage, the AI storm is shaping up as the next chapter in that argument, and the stakes have rarely been higher.
Sources: The Australian Financial Review.
