Reflections from the 6D AI Melbourne CXO Leaders Panel
For the second year running, Intelligent Pathways sponsored 6D AI Brisbane. This year’s conversations felt different. Last year, much of the focus was on what AI could do. This year, the conversation shifted to something more practical. How AI actually shows up in real work. How it changes decision-making and what it means for organisations trying to move beyond experimentation.
I had the opportunity to join the panel on The Age of Agents: The Future of Work. While the topic suggests a move towards fully autonomous systems, what came through clearly from the discussion is that the reality is more grounded. We are not moving towards fully agentic organisations. We are moving towards better orchestration of work that increasingly involves agents.
The Gap Between Concept and Reality
There’s a lot of momentum behind agentic AI at the moment. The idea of chaining agents together, giving them objectives, and letting them run end-to-end. It’s a compelling concept. But when you start applying it to real environments, things get messy pretty quickly.
One of the other panellists, working in financial services, described it as putting a brick on the accelerator and calling it autonomous. It works for a short burst, but it’s not something you’d trust over the full journey.
That resonated. Don’t get me wrong, autonomy is very real and beneficial in many scenarios but when the road isn’t always straight we need to understand when to pump the brakes and push the accelerator.
In most of the environments we work in, processes aren’t neat. They don’t start and finish in a single interaction. A claim can run for months. A healthcare scenario can shift minute by minute. Engineering decisions can carry consequences for decades. Trying to force those into a fully agentic model introduces risk in addition to value. What’s emerging instead is a more balanced approach. Some parts of the workflow are deterministic, some benefit from agentic capability, and some still rely on human judgement. The challenge is knowing where each belongs and finding that perfect balance.
Designing for Decisions, Not Processes
One of the more useful ways to think about this is to step back from the idea of automation altogether.
Start with the decision.
- What are we trying to resolve?
- What information is needed?
- What does a good outcome look like?
From there, you can design how AI supports that decision.
I shared an example from an emergency healthcare setting. You’ve got a triple zero call coming in, fragmented information, and clinicians trying to make sense of it quickly. What we’re introducing are agents that sit alongside that process. They’re not making final decisions, and they’re not replacing the clinician. They work through the information as it comes in. Extracting clinical details, interpreting what’s being said, and mapping that against known protocols to bring to the surface the information that supports the best decisions for a clinician.
In one case, identifying something like a child swallowing a button battery and immediately surfacing the relevant clinical pathway and critical time windows. They also continue to observe the situation. As treatment progresses, they track how things are evolving against those windows and whether the recommended approach shifts. All of that happens in the background.
By the time a clinician engages, they’re not searching through documents or trying to recall edge-case protocols under pressure. The system has already pulled together the relevant information and presented it in context. The role of the clinician doesn’t disappear. If anything, it becomes more focused. Less time navigating information, more time applying judgement where it matters.
Orchestration Is Where the Value Sits
This idea came up repeatedly across the panel, even if we didn’t all use the same language. The real work isn’t building agents. It’s designing how they fit into a broader system.
A technology leader from the insurance sector spoke about this in terms of how you bring a human into the loop. Not asking them to manage the entire process, but designing a moment where they step in with the right context, make a decision, and then hand back to the system. That shift is significant. Historically, we’ve designed systems where people carry the process. Now we’re starting to design systems that carry the process and involve people when it matters.
Another panellist pointed out that if you zoom in far enough, everything can be automated. But when you zoom out, you still see augmentation. The system might handle more of the work, but the human role doesn’t disappear. It moves. That’s what orchestration looks like in practice.
Trust Doesn’t Come From the Technology
There was also a lot of discussion around trust. Whether people are ready to rely on these systems. Whether AI needs to be “better than humans” before it can be used in critical environments. I don’t think that’s where the real issue sits. In the healthcare example, clinicians aren’t questioning whether AI should be there. They’re focused on whether it helps them do their job. If the system is clear, if it shows its reasoning, and if they can challenge it, they’ll use it.
One of the panellists made the point that in many narrow areas, AI is already outperforming humans. At the same time, there are situations where human intuition still plays a critical role. Particularly when something unexpected happens and there isn’t a clear pattern to follow. It’s not a competition. It’s about understanding how the two work together.
The Shift in Roles Is Already Happening
Naturally, the panel conversation turned to how roles are changing. One of the panellists framed it as a shift from “bricklayers” to “architects”. It’s a simple way of describing what’s happening across most industries. Work that is repetitive and well-defined becomes easier to automate, while the value shifts to roles that operate at a higher level. Designing, coordinating, and thinking about how things fit together. Someone pointed out the irony – that in the real world, bricklayers, along with plumbers and electricians, are probably among the safest jobs right now. Which got a laugh, and a quick reminder not to break the metaphor. But the point still holds.
Separately to that, I shared how this shift is playing out in software engineering. As developers, you’ve always had a mental model of what you’re building. You picture the city as you go. The streets, the buildings, how everything connects, and over time you get better at expressing that as code. That thinking doesn’t change. What changes is how you bring it into existence. Instead of building each part directly, you’re describing the city. You’re articulating what needs to exist, how it should behave, and how it fits together. If that description is clear enough, the system can construct it with you, often very quickly.
So the role isn’t disappearing. It’s shifting. You move from building each piece to clearly defining the outcome. And that’s where we’re seeing people start to get the most value from these systems.
What Actually Creates Advantage
Towards the end of the panel, the question came up around competitive advantage. If everyone has access to the same tools, what makes the difference? The answer wasn’t technology. It was how organisations work. How well business and technology are integrated. How quickly teams can move from an idea to something real. How effectively they can apply these tools to their own context.
Most organisations are capable of doing the same things. Our experience shows the difference comes down to people, culture, and leadership.
Where This Leaves Us
If there’s one thing I’d take from the discussion, it’s that we need to be careful about how we frame the future of AI. It’s easy to focus on autonomy. On what the technology might eventually do. But the more immediate challenge is designing how work actually happens. Where decisions sit, how context is shared and when people are involved. The organisations that get this right won’t be the ones chasing fully agentic systems. They’ll be the ones that understand how to combine them with human judgement in a way that holds up in the real world. That’s where we’re focusing our time. Because ultimately, this isn’t about the agents. It’s about how you use them.
If you're ready to explore AI’s strategic potential for your organisation, let's connect.
Schedule a Strategic AI Discovery Session today.
