02Sanjay
An installation already knows what happened. It just can’t say so yet.
Sanjay connects structured data, unstructured information, real-time signals and geospatial context into a queryable intelligence layer and living digital twin.
Explore in 3DThe problem
Take a working installation: aircraft, vehicles, sensors and people, operating normally. Everything on it is already described somewhere, by four kinds of information.
Every asset is already producing data.
Telemetry, radar tracks, link quality and conditions, streaming from the equipment itself.
Systems of record describe the same assets.
Asset registers, maintenance systems and mission databases.
Reports and briefs hold what databases don’t.
Assets, dates, units and places, extracted from the documents people actually write.
Geography ties it to the ground.
Sectors, routes, coverage and coordinates, in the same model as everything else.
Each world is complete on its own terms, and none is aware of the others. The same vehicle, VEH-2047, exists six times across six systems with six spellings, and no system knows they describe one vehicle.
So when something happens, the information exists but the context is scattered. Answering one question means eight teams opening eight systems and assembling the picture by hand:
How Sanjay works
Sanjay sits across the systems an installation already runs. Nothing is migrated or replaced; the sources are connected where they live and modelled as one.
- 01Connect
Connect every source. Replace none.
Databases, documents, sensors, GIS and external feeds, connected where they already live.
- 02Entity resolution
Six records. One vehicle.
Every record that refers to VEH-2047 becomes one entity, with its sources kept.
- 03Ontology
Entities, and how they relate.
Assets, units, missions, places, events and documents, in one connected model.
- 04Ask
Ask anything about your operations.
A plain-language question is traced through the model, across every system that holds part of the answer.
- 05Evidence
Every finding traced to its source.
Answers come back as findings, each with the records and systems it was drawn from.
- 06Digital twin
The model, back on the ground.
Every entity bound to the asset, place or event it describes, and kept current.
An investigation, end to end
At 19:42 in Sector North, something is happening, but no single signal proves what. Five observations arrive in four systems, eighteen minutes apart.
- 19:31:40Unit BravoEnters Sector North
- 19:42:18S-17Unusual movement detected
- 19:47:05COMMS-08Signal drops to −91 dBm
- 20:05:12VEH-2047Leaves planned route, 1.2 km
- 20:09:40RADAR-17Low-altitude track #1847
- 03 JulArchiveSITREP mentions comparable activity
Sanjay traces the question through the model, from the sector’s GIS polygon to the vehicles, units, radar tracks, sensors, comms events, documents, maintenance records and mission window that touch it, and returns potential contributing factors:
- 01
An abnormal communication event occurred near the sector.
19:47:05COMMS-08 · signal logEVENT #421GIS · Sector North - 02
VEH-2047 changed route approximately 18 minutes later.
20:05:12Telemetry · VEH-2047GIS · planned route R-4 - 03
S-17 recorded unusual movement in the same area.
19:42:18SENSOR S-17 · detection logGRID 18S 4208 3176 - 04
A previous report references comparable activity.
03 JulMaintenance report.pdfSITREP · Sector NorthAnalyst note - 05
Historical records show similar activity near the same geographic region.
03 JulArchive · EVENT #910.6 km from EVENT #421Sector North
Confidence · Moderate (0.64). Correlated observations, not a confirmed cause. 5 findings, 7 source records, 4 systems.
One context
One environment. One context.
The model goes back onto the ground as a digital twin: every entity bound to the asset, place or event it describes, and kept current. The walkthrough shows all of this happening on a live 3D installation.