Context: International team project · XRCC
My role: Product and technical direction, team coordination and hands-on development
Status: XRCC finalist · Prototype in active development
The problem
Property claim inspections require people to move between checklists, photos, notes and reporting tools while examining a property. InVi AI explores whether a heads-up workflow could keep guidance and evidence capture closer to the inspector’s task.
My role
I lead the product and technical direction and contribute hands-on to development. My responsibilities include coordinating an international team across Canada, the USA, India and Australia, shaping the prototype architecture and user experience, developing the companion app experience, and leading XR and AI integration.
Test the workflow before committing to hardware
At the proof-of-concept stage, we did not have access to the target wearable display hardware. We used Quest 3 to explore the inspection flow through a lightweight heads-up interface, rather than building an experience dependent on immersive VR.
That gave the team a way to work through the interaction model while access to the intended devices was still a constraint.
Make the prototype trade-off explicit
When we gained access to Meta Ray-Ban Display during the hackathon, we had to choose between native and web-based approaches. The native route provided camera and microphone access but a more limited visual experience. The web route supported a stronger interface, without the inputs needed for the complete workflow.
As a team, we prioritised demonstrating the wearable interface and overall product experience. We used the web interface on the glasses and simulated the missing camera and microphone inputs through the companion mobile device.
That was a deliberate prototype decision. It helped us demonstrate the intended workflow, while leaving deeper native integration as a requirement for a production implementation.
Outcome and next steps
InVi AI was selected as an XRCC finalist and remains a prototype in active development. The work demonstrates a product direction and a set of platform decisions, rather than a production-ready insurance service.
Before this could be treated as a production service, the complete workflow would need validation on target hardware and evaluation with real users. Faster inspections, improved consistency and reduced administration remain goals to test, not measured outcomes claimed here.