The Spatiotemporal Intelligence Automation Plane

The Spatiotemporal Intelligence Automation Plane

Physical AI is the new layer of the AI stack: software that does not just predict, but acts on the physical world — through agents, robots, machines, and humans. To act on the physical world, that software needs more than models. It needs a substrate that connects each physical Thing to whatever is acting on it.

That substrate is the Spatiotemporal Intelligence Automation Plane, or SIAP.

Where the SIAP sits in the Physical AI stack

A Physical AI stack has three layers:

The plane is the middle layer. It turns a forest of disconnected Things into one spatiotemporal substrate that every actor in the stack can reason about.

What the SIAP actually does

A SIAP combines three inputs into a single operational plane:

Every event is anchored to a When, a What, and a Where, in that order. The plane keeps the full history — not just the last ping — so any actor can reason about *where a Thing has been*, not only *where it is*.

Every Thing connected to four actors

The plane has no preferred actor. Whatever is acting on a Thing — software or steel, autonomous or operator-driven — talks to the Thing through the same spatiotemporal interface.

The substrate: sustainable radiofrequency unique identifiers

For the plane to connect Every Thing, every Thing needs an identity it can carry. We use sustainable radiofrequency unique identifiers — passive or RF-rechargeable, no shipped lithium cells, long shelf life, item-level cost, and readable by the same RAIN UHF infrastructure already deployed in retail, logistics, and food operations.

Spatiotemporal Intelligence, Automation — what each term means

How a SIAP works in practice

A SIAP is built from several cooperating elements:

1. Sustainable RF identity at the Thing level. RAIN RFID and related technologies give each Thing a persistent identity that can be sensed repeatedly at scale.

2. Event capture at the edge. Fixed readers, handheld readers, mobile devices, and operator workflows create timestamped events as Things move.

3. Spatial context. Each event is paired with a location identifier or inherited GPS coordinates so the record is spatial as well as temporal.

4. Reconciliation with enterprise systems. Events are matched against ERP, WMS, planning, and workflow systems to detect drift between reported state and actual state.

5. Automation loops. Agents, robots, machines, and humans subscribe through REST and MCP APIs. The plane triggers alerts, pricing decisions, reallocation, exception handling, and AI model updates — in time.

Why this is more than asset tracking

Traditional asset tracking answers one narrow question: where is the thing right now?

The SIAP answers a broader and more valuable set of questions:

The value is not in seeing a dot on a map. It is in understanding delay, dwell, sequence, and consequence — and surfacing that to the right actor.

Cooldat®: an instance of the plane

Cooldat® is what the SIAP looks like when it is wired up for perishable supply chains: CoolTag sensors on every item, QDatDroid and QDatFX reader clients at the edge, and QDat.io in the cloud — a single end-to-end instance where agents, robots, machines, and humans share the same per-item temperature truth.

That same data feeds predictive shelf-life models, which can be executed at the edge through QDatDroid to support faster and more autonomous decisions in cold storage and other complex environments.

Why Physical AI needs the SIAP now

Physical AI already exists in pieces. What is missing is the plane that makes the pieces coherent.

Without a SIAP, agents, robots, machines, and humans each see a different slice of reality. With a SIAP, they all see the same Thing-level history — anchored in When, What, and Where.

That distinction becomes critical wherever value decays with time, where inventory can appear valid while the field says otherwise, and where AI systems need trustworthy operational evidence rather than delayed summaries.

Getting started

A SIAP does not begin with abstract architecture. It begins with one Thing flow, one identity model, and one repeated operational blind spot.

From there, the system expands: more readers, more locations, more Things, more reconciliations, and more actions driven from reality — by every actor on the plane.