Construction sites, cities, infrastructure and real-estate assets now generate enormous volumes of data. The challenge is no longer collecting information—it is turning information into understanding, and understanding into action.
Observe — understand what is happening on site
The first step is gathering evidence from drone photography, orthophotos, photogrammetric models, LiDAR, point clouds, 360° images, BIM, GIS, documents, measurements, sensors and earlier surveys. Each source reveals a different part of reality, but collecting files alone does not reveal what requires attention.
Orient — turn data into context
If AI detects a polygon in an aerial image, the detection alone is not enough. The system needs to know whether it is a building, which parcel and address it belongs to, whether it existed in the previous survey, which road is nearby and which open tasks or BIM elements relate to it.
An entity-first approach treats a building as a persistent digital entity with identity, location, geometry, history, imagery, documents, tasks, planning information and AI results. The question changes from “what is in this file?” to “what do we know about this building?”
From layers to relationships
When buildings, parcels, roads, captures, tasks and changes are connected over time, the system forms a Spatial Knowledge Graph. It holds not only geographic data, but also an understanding of how things in the real world relate to one another.
Decide — move from viewing to decisions
With observations and context, the system can identify new construction, demolition, roof changes, earthworks, volume changes, roads, vegetation or equipment. It can combine geometry, AI confidence, open tasks and missing evidence to decide which area requires human review.
The same context can guide the next capture: which facades have already been documented, what is occluded, and where information is still missing. This is where a digital twin becomes a decision tool rather than only an impressive model.
Act — turn a decision into work
A detected roof change can create an inspection item. Missing facade coverage can produce a capture plan. A possible deviation can open a task. A completed survey can generate a change report. The system moves from “here is the data” to “here is what happened, why it matters and what should happen next.”
Action becomes the next observation
OODA is a loop. Every action produces new evidence, and that evidence begins another cycle. The digital representation becomes a living, continuously updated memory—not a snapshot of one moment.
From a 3D viewer to a Spatial Operating System
A viewer asks what can be seen. GIS asks what information is here. A digital twin adds the current digital state. A Spatial Operating System goes further: what is happening, what changed, why it matters, what is missing and what should happen now?
Skylens is being built as a persistent spatial memory of the physical world—one that people, software systems and AI agents can observe, understand, decide through and act upon.