Give objects an identity
Extract and frame assets from the supplied scene so equipment can be selected and understood in context.
Since AI 2026 / 6–8 Nov / Turku
A room becomes more useful when you can understand what is in it, find the right equipment, and know what to inspect next.
The idea
Industrial context is scattered across physical spaces, device readings and maintenance documents. Scan2Work brings those pieces into a shared 3D environment.
Start with the supplied 3D Gaussian-splat capture of an industrial boiler room.
The physical contextFrame individual assets and connect them to synthetic equipment records, conditions and telemetry.
The equipment contextFollow a technician’s route, surface nearby equipment and review evidence for the next inspection step.
The service contextThe Valmet Automation challenge
Built at the Since AI 2026 hackathon in Turku, Finland, 6–8 November 2026, for Valmet Automation’s challenge.
About the hackathonExtract and frame assets from the supplied scene so equipment can be selected and understood in context.
Map the provided synthetic Excel equipment, telemetry, alarm and service data to the extracted assets.
Simulate obstacle-aware technician navigation and show nearby assets as the technician moves through the room.
Use device conditions and supplied troubleshooting documentation to support diagnostics with traceable evidence.
The people behind the prototype
Five people, one shared challenge. Meet the team that came together to explore what industrial spatial intelligence could become.
Explore the prototype