UX Australia 2026 — Day 2, Session 2

Designing for the Human Margin is a talk about the precise, mappable territory AI cannot take from design — and why our discipline has been quietly giving it away for twenty years.

For decades, we've designed technology to help people react.

An alert appears. A dashboard turns red. Teams scramble to investigate, triage, and resolve problems after they've already happened.

But what if technology could help prevent problems before they even begin?

As AI becomes embedded into the products we build, UX is undergoing a fundamental shift. Our role is no longer just to design interfaces or simplify workflows. We are designing intelligent ecosystems that understand context, learn behaviors, and proactively help people achieve their goals.

In this talk, I'll share my journey designing AI-powered experiences for complex enterprise environments and the framework that emerged from that work. Rather than adding another AI assistant to the interface, we'll explore how intelligence can quietly work in the background, reduce cognitive load, and help people focus on what matters most.

The future of UX isn't about building smarter interfaces. It's about designing experiences that help people make better decisions before they even realize they need help.

Because the best technology is deeply human.

User experience design often focuses on usability, accessibility, and task success. But even when users complete a task successfully, an interface may still be demanding unnecessary mental effort.

Every extra decision, notification, unclear hierarchy, or fragmented workflow adds to the cognitive load users must manage. These costs are rarely visible in traditional usability metrics, yet they directly affect attention, performance, and user satisfaction.

Cognitive Load Theory offers a way to make this invisible cost visible.

In this talk, Dr Ben Shelton and Cordelia Prangley introduce how cognitive load can be used as a practical lens for evaluating interface effectiveness. Drawing on research from educational psychology and calm technology, the session explores how mental workload influences user performance, decision-making, and attention.

Attendees will learn:

- How cognitive load affects user performance and decision quality - The relationship between perceptual load, attention, and interface complexity - Practical techniques for identifying cognitive overload in real products - How cognitive load insights can lead to calmer, more effective digital experiences

If we want technology to feel effortless, we need to understand the demands it places on the human mind.

Unplanned disruption is where complex systems are most exposed. In transport, it brings together multiple modes, fragmented systems, and teams making decisions quickly with incomplete information. The result is not just operational pressure, but visible gaps in governance, processes, tools, and communication that directly impact the customer experience. At the same time, it creates challenges for both customers and employees, with frustration often building between teams that are each working within different constraints and realities.

This session explores how human-centred design was used to make those gaps visible and actionable. Drawing on practical work across cross-agency collaboration, standardised processes, and shared tools, it shows how mapping the disruption ecosystem, visualising customer and staff journeys, and establishing a shared definition of disruption can support clearer governance, more consistent decision-making, and better coordination in real-time. This program of work helps stakeholders see the broader system they are operating within and reveal the experiences of others across it. This created empathy across teams, reduced some of the frustration that had built up between functions, and gave people a clearer sense of why their work mattered in the bigger picture.

Attendees will leave with practical approaches for mapping complex systems, aligning cross-functional teams, and improving decision-making in high-pressure, real-time environments.