From Idea to Reality: Lessons from Turning Research Projects Into Enduring Products

Over roughly a decade working in research commercialisation and technology, I have taken many different projects through the journey from idea to a working product or company. Each one reached some version of the market. Each one taught me something different. This post covers three of these, and two out of three I would genuinely call successful.

Two out of three is a good hit rate as far as I’m concerned. Research and technology commercialisation is not a process you can run reliably from a template. But there are patterns in what works and what does not, and the most important of them start long before anyone writes a line of code or files for a patent.

Here is what I learned from each.

Project One: Envision

Status: Housing Policy. Funding: Grants. Lesson: Focus on the right problem.

Envision started from a real and genuinely important insight. When middle-ring suburbs of cities are redeveloped, the outcome is almost always worse than it could be. Land gets locked up for another fifty years in a form that does not reflect what the community actually needs. Local governments have limited tools to identify which sites could be better used, or to have productive conversations with developers about what better development might look like.

The problem was real. The people who had the problem were real. State government, local government, developers, community groups. We built an online tool to help councils find sites suited to greyfield development, where ageing housing sits on land whose value and location suggest a better use is possible.

The product worked. It still exists. But I was not deeply satisfied with the outcome. The reason was that we had not been precise enough about which of those stakeholders would actually pay to solve the problem. State government was interested, until they were not. Councils were willing to pay, but their budgets are constrained. Developers had no incentive to use a tool that identified sites they had not chosen themselves. The funding model was grant-dependent, which meant lumpy development cycles, gaps in use, and no strong market signal.

The lesson was not that the problem was wrong, but that we had not been disciplined enough about who would pay to solve it, and we had moved to building the solution before that question was properly answered. In the end, the problem that needed solving was one of policy and community engagement, not one of technology. It is this policy and engagement effort that continues in Swinburne University and Maroondah Council to this day.

Project Two: Value Australia

Status: Operating company within PEXA. Funding: Collaborative Investment Lesson: Timing is as important as planning.

Value Australia is the one I would point to if someone asked for a genuine research commercialisation success story. It is also the one where, if I am honest, the timing of certain decisions and relationships was as important as the quality of the work.

The project was a collaboration between FrontierSI, UNSW, Commonwealth Bank, NSW Government, and several partners. The goal was to build Australia's first highly accurate, transparent, and interactive automated property valuation model, one that could tell you what a property was worth, explain why, and let you model what would happen to the value under different scenarios.

The problem being solved was significant. Property valuation in Australia is expensive, inconsistent, and opaque. Planning decisions, infrastructure investment, lending, insurance, and tax are all affected by valuations, and the existing models were not fit for the decisions being made with them.

After three years of development we had commercially valuable IP and a clear sense that there was a market for it. We worked through the commercialisation path deliberately: engaging the right advisors to validate product-market fit, developing a go-to-market strategy, pitching to strategic partners, surviving a due diligence process that involved fifteen people on the other side and hundreds of pages of agreements, and ultimately establishing a company called Slate Analytics, with PEXA as the strategic partner and majority owner.

The deal returned millions to FrontierSI and UNSW over a four-year period. It was later recognised as a Research Commercialisation Award winner by Cooperative Research Australia, described by a specialist lawyer as a remarkable success rarely seen in collaborative research programs. 

But timing was a genuine factor. The right strategic partner appeared at the right time with the right strategic rationale. The deal stumbled multiple times, including a significant tax challenge before signing! COVID delayed everything. The lesson isn’t that planning does not matter. It is that even when you do everything right, you need the environment to cooperate, and you need to be in the market long enough for the right opportunity to find you.

Project Three: MapAI

Status: Operating start-up. Funding: Shareholder Investment Lesson: Get the right support early.

MapAI was the most deliberate of the three. Having been through the previous two experiences, I came to it with a clearer sense of what I wanted to do differently.

The problem was real and well understood: good decision-making is held back by a lack of easy, quick access to insights locked inside geospatial data. Most organisations with rich spatial data sets cannot easily query them, cannot surface patterns without significant analyst time, and certainly cannot let non-technical staff ask questions and get useful answers. Combining conversational generative AI with geospatial analysis in a transparent, explainable way addressed this directly.

This time we spun the company out from the start with a clear intention to build for investment. We raised pre-seed funding, developed the business strategy and product roadmap with board oversight, and raised a further seed investment, reaching a multi-million post-money valuation within six months of launch. We secured foundational customers, built strategic partnerships, and won a core role in a $6.5 million, three-year project to develop an AI urban planning product in partnership with Archistar and PEXA.

The lesson from MapAI was about getting the right support around you from day one: the right investors, the right advisers, the right board, and not waiting until you need them desperately to find them.

The Pattern Across All Three

Looking back across these three experiences, the framework that connects them is not complicated, but it requires discipline to follow in order.

Understand the problem first, specifically, in the language of the person who has it. Find out who has the problem and whether they will pay to solve it. Only then design the solution. Choose your first customers carefully and build with them, not for a general market. And think hard about the sustainability model before you are in crisis about it.

The places where I have seen this process go wrong are always the same. Teams skip the early problem validation and go straight to building. They confuse interest from potential users with willingness to pay. They build for everyone and end up serving no one well. They leave the question of what happens after the project until the project is almost over.

 None of this is new advice. But in my experience it is rarely followed consistently, particularly in research and technology environments where the technical work is genuinely exciting and the commercial work feels like a distraction. It is not a distraction. It is the foundation that determines whether the technical work ever has any impact at all.