Beyond customer attitudes
The approach needed to capture observable behaviours and the contexts influencing customers' willingness to delegate decisions to AI.
As part of a multidisciplinary proposal developed by a consulting consortium, I designed a framework connecting behavioural research with the structured data, decision rules and interaction logic required to inform an evidence-based human simulator.
The framework established a coherent approach across behavioural science, design, data and AI decision modelling.
A research-to-system framework integrated into the consortium's final client proposal.
I developed the framework’s principles, output structures, connected phases and worked examples, and contributed to the proposed Decision Model and behavioural archetypes.
External consulting — proposal development. The framework was incorporated into the final proposal, which was submitted and was not awarded.
A major Australian financial institution was exploring how customer behaviour, trust and decision-making could inform AI-powered banking interactions. The challenge was to move beyond understanding customer attitudes and define how research evidence could inform the decisions AI agents make.
The approach needed to capture observable behaviours and the contexts influencing customers' willingness to delegate decisions to AI.
Traditional research outputs would require further translation before they could inform technical requirements and decision models.
The proposed simulator needed logic to distinguish when an agent could act, when customer confirmation was required and when to hand control to a human.
Behavioural science, research, design, data and AI expertise needed to operate as a connected process rather than separate workstreams.
I developed the principles, output structures and connected phases of a research-to-system framework, showing how behavioural observations could progressively become structured inputs for AI decision modelling.
I established principles that positioned behavioural research as a source of structured evidence for downstream systems, not solely as an insight-generation exercise.
I defined output types connecting behavioural observations to structured records, behavioural patterns, decision rules and AI behaviour instructions.
I organised the proposed approach so each phase would generate outputs that could serve as inputs for the next, aligning research activities with downstream technical needs.
I developed worked examples illustrating how customer context, behavioural signals and decision conditions could be expressed as system rules.
I helped define the proposed execution, approval and escalation logic, alongside behavioural archetypes designed to inform different interaction requirements.
Developed a framework linking behavioural research, structured data, behavioural signals and AI decision modelling, establishing clear relationships between the proposed workstreams.
Created output definitions and worked examples to demonstrate how research evidence could be converted into structured, actionable inputs for technical teams.
The framework was incorporated into the consortium's final client proposal, informing the approach, data structures and behavioural decision model.
The work demonstrated how a complex, multidisciplinary research proposition could be structured around practical outputs, connecting what an organisation learns about people with what its systems need to do.