Jordi Spranger describes a gap between financial building records and the changing physical state of a space. His Spranic demonstration uses a handheld depth sensor to capture a room, identify objects and organize them into a semantic graph. The aim is a shared, queryable memory that people and machines can revisit rather than an isolated map that disappears after a robot leaves.
Jordi Spranger compares two kitchen captures to show how observations can be aligned across time. Natural-language questions locate objects, while linked capture evidence supports claims that chairs have moved. Keeping observations and their time context makes it possible to inspect the basis for a change report rather than treating a generated description as ground truth.
Jordi Spranger proposes applications in maintenance, inspections, insurance, interior design and robotics. These are possible uses of the approach, not demonstrated premium reductions, certified compliance or independently measured commercial benefits. Accurate captures, current records and validation remain necessary before a spatial graph can support consequential decisions.
Watch on YouTube
