TwinMS Intelligence

From operational data to operational understanding.

Intelligence in TwinMS is not a chatbot, an orb, or a dashboard. It is the layer that gives every signal a place in the physical world, learns what normal means there, detects what matters, predicts what comes next, and acts. This page shows how one signal becomes understanding.

The life of a signal

Watch one signal become understanding.

One real reading, followed through the whole intelligence cycle: from raw telemetry to a completed work order and a smarter baseline. Select a stage, or let it play.

It begins as signals.

216+ sensor channels stream air, energy, water, occupancy and access on a 10-second cadence, hydrated into the platform in under 50 milliseconds. Alone, these are just numbers — accurate, fast, and meaningless without a world to belong to.

The twin gives every number a place.

Each reading arrives room-aware, floor-aware, asset-aware and time-stamped, anchored to the digital twin built from the ISO 19650 model. A temperature is no longer a value; it is Main Hall, Level 4, beside AHU-04, at 15:40:31.

It learns what normal means here.

730 days of history build living baselines per system and per season: how this building breathes on a summer Thursday, what this pump draws under load. Understanding is not a threshold someone typed in; it is learned operational reality.

Anomalies you can trust.

When behavior diverges from its learned baseline, TwinMS flags it, with baseline recomputation continuously suppressing false positives. When the platform raises its hand, it matters.

It sees the fault before the failure.

Trends project forward: predictive wear-scoring on critical assets and anomaly trajectories flag likely faults days ahead with a confidence you can act on, in time to act cheaply.

Your rules stay in command.

A configurable rule hierarchy — sensor → space → floor → type → global — decides what runs automatically and what waits for human approval. Intelligence proposes; your operating policy disposes.

The loop closes.

Work orders dispatch to CMMS, BMS setpoints adjust, field teams get it on mobile, cameras jump to the event. The prediction becomes a completed job, without a human copying data between systems.

Every outcome makes it smarter.

The result — resolved, rebalanced, verified — is written back into the record. Baselines recompute, wear models update, and the platform understands the operation a little better than it did yesterday.

Two kinds of memory

Real-time speed. Historical depth.

The last ten seconds

Live telemetry on a 10-second cadence, hydrated in under 50ms: the operation as it is right now, not as it was in last month's report.

The last two years

730 days of history per system and season, the memory that turns a reading into a judgment: is this normal here, now, for this asset?

Combined, they judge

Every live value is scored against its learned past. That combination, not a static threshold, is what makes TwinMS alerts worth trusting.

Context from the twin

Intelligence needs a world.

Generic AI sees numbers. TwinMS intelligence sees the building, because the digital twin gives every data point spatial, asset and enterprise context. That context is what turns detection into diagnosis: not "sensor 114 is high," but "the meeting room beside AHU-04 is overheating while the unit's draw is climbing."

Room-awareFloor-awareAsset-awareSystem-awareTime-stampedTicket-linked
Human + AI

Operators stay in command.

Automate what you trustRoutine responses — setpoint corrections, standard work orders — run automatically under your rules.
Approve what mattersAnything outside policy waits for a human decision, with the full context attached, in the hub or on mobile.
Audit everythingEvery reading is time-stamped and room-aware; every action is logged from detection to verified fix.
Rule hierarchySensorSpaceFloorTypeGlobal
Within policy · setpoint drift, routine service
runs automatically
Outside policy · critical systems, cost thresholds
waits for approval
Governed intelligence

Trusted, because it's accountable.

No black boxes: alerts trace to the readings and baselines that raised them, actions trace to the rules that allowed them, and access is governed by enterprise sign-on. Built in the Kingdom, aligned with its mandates.

Full audit trailSAML 2.0 + MFARole-based accessFirst Saudi-builtPDPL-ready

See your operation, alive.

A live demonstration on a real building: your systems, your scale, your questions.