Troy Neilson (Graduate Certificate Business Admin ’20; MBA ’21; PhD ’24) (MBA, Technology Strategy) has spent the past two decades working in AI. But even for him, every project still brings so many unknowns. “The tools that kicked off today’s AI revolution just three or four years ago aren’t anything like what we use today,” he says. “The evolution of AI as a field is one of the most rapid advances we’ve seen for a long time. In a short period, we’ve gone from zero to hero.”
Troy’s career has seen him design, build and deploy more than 100 machine learning and generative AI models across financial services, agriculture, biotech, life sciences and enterprise. Today he focuses on creating agentic AI systems that reason, plan, transact and reconcile. He’s currently CTO and co-founder of Glassbox Labs, an AI company with three functions: carrying out applied engineering for clients, creating its own products, and pursuing a small research program.
In this space, the only thing that’s certain is uncertainty. So when Glassbox Labs was asked to work with a global company to build agentic business-to-business transactions for SMEs—enabling them to pay and get paid more efficiently—he knew it would be a challenge.
“The world is shifting very quickly, and so even large organisations with head counts in the tens of thousands don’t have the specialised knowledge that they need,” he says. “They lean on people like us to help them push forward.”
The first application that the client built had run into difficulties. “Literally, on the day of launch, there were a lot of struggles and drama. They were still iterating over this application on the day they were looking to launch it.” In the first conversation with Troy, the client asked if he could build nine versions. “I said that I didn’t even know if we could build one!”
This particular piece of technology had never been constructed before, so Troy didn’t have a reference case. But that wasn’t even the main problem. “First, I didn’t have an understanding of what needed to be built. And the other big difficulty was that the client didn’t have an understanding of the problem.”
That’s what happens when you’re doing something that’s so close to the edge of new, he points out. “Your client doesn’t know what good looks like, because they haven’t yet had any of their customers come back and tell them what they need or want, or how to fix a problem.”
Nobody could give Troy and his team definitive answers. But they had to move forward. “The only thing you can fall back on is what you know and the only thing you can control is your response to things. Accepting those facts goes a long way,” Troy says. So he and his team started by making assumptions around what good and bad looked like. “We had to define problems that didn’t exist and then build solutions to them.
Plus, we had to build different use cases across two different countries, contending with different cultures.”
The key to what might seem a highly technical problem was in understanding human connection. “From a technical perspective, everything is kind of solvable. But when it comes to humans, it’s a lot more difficult.” In fact, that was the hardest part for him, he says. “I’m a tech nerd, so applying human elements and creating a meaningful customer journey was tricky.”
They had to think about every step from someone else’s perspective. “How are people going to use it? What problems are they going to encounter? What about accessibility or language?
“How will this work for very different businesses who are all trying to deliver value for their customers? How is this going to solve a real problem for a real person? If it doesn’t solve a real problem for a real person, it’s meaningless.”
And the approach worked. The new project is now up and running successfully and Troy is now eager to embrace the next big opportunity. From an innovation perspective, he says, Australia is leading the AI pack. “We have a unique market in the sense that we have the talent, we have the capability and we tend to be first movers.
“It’s so exciting. It’s such a massive space. Every day you’ll see some other application or tool or use of AI that you’ll never have thought of, and you’ll think: that’s so cool.” In his position as Adjunct Research Fellow at Griffith, he’s also bringing experiences like this into how to help students apply AI in their fields.
Troy has been in the startup game for 25 years. He knows how it feels to face a problem that seems insurmountable. And what’s helped him the most, he says, is the ability to get up again after being knocked down. “If you know that this is the right thing to do, or if it’s the thing that you love doing, then it’s about being able to pick yourself up every single day and keep moving."
Want to skill up wherever and whenever you like? Check out our 100% online, on-demand degrees
Image captions (top to bottom):
- Troy Neilson.
Published 29 October 2026
Explore more from the newsletter
Sustainable Development Goals
Griffith University is aligned to the United Nation's Sustainable Development Goals (SDGs) and is committed to promoting quality education, decent work, and industry, innovation and infrastructure, working towards a world where innovation drives sustainable development.
Stay connected
Keep in touch with your alumni community.
