Asset Agents · Persistent engineering context | ENGINPILOT
PLATFORM / 06 — ASSET AGENTS

One agent per asset. It remembers what happened in 2024.

An asset agent is a persistent, addressable entity bound to a single physical asset. It holds that asset's design basis, its operating history, its failure modes and its open exceptions — and it proposes. It does not decide, and it cannot act.

STATUS IN DEVELOPMENT · ARCHITECTURE DEFINED · NOT YET DEPLOYED
SCOPE
One asset, one agent, one accumulating context
GROUNDING
The graph, not raw documents
OUTPUT
Proposals with sources, confidence and verdict
AUTHORITY
None — a named engineer signs
01 — WHAT AN ASSET AGENT IS

Addressable, persistent, bound to one asset

A general assistant is asked a question and forgets it. An asset agent is a standing entity in the graph: CH-02 has an agent, and that agent carries the chiller's design basis, its commissioning data, every intervention since, and the exceptions still open against it.

This matters because engineering context is cumulative. The reason a bearing failed in 2024 constrains what a sensible recommendation looks like in 2026 — and a stateless model has no way to know that.

IDENTITYOne agent per asset tag, for the asset's life
MEMORYEvery proposal it made, and whether it was accepted
HANDOVERSurvives the engineer who commissioned the asset
AGENT CONTEXT · ACCUMULATED ASSET CH-02 CHILLER · 1250 kW COMMISSIONED 2019 ASSET AGENT 2019 NOW DESIGN BASIS FMEA REV B BEARING FAILURE2024 · ROOT CAUSE HELD RETROFIT OPEN EXCEPTION A STATELESS MODEL SEES ONLY THE LAST COLUMN THE AGENT SEES THE WHOLE RECORD FIG 1 / ACCUMULATED ASSET CONTEXT · NTS EP-AGT-001
02 — WHAT IT CAN DO

Four capabilities, invoked on its own asset

The agent is the addressable entity; these are the functions it runs. Each operates over the graph rather than raw documents, and each cites the nodes it used.

◆ RETRIEVAL

Knowledge

Answers engineering questions about this asset with citations to the exact document revision and clause.

◆ DIAGNOSIS

Failure mode

Ranks candidate failure modes against observed symptoms and this asset's own FMEA — never a generic library.

◆ PLANNING

Intervention

Proposes inspection and maintenance timing, every option passed through PhysicsNET first.

◆ REVIEW

Compliance

Checks that decisions and their evidence satisfy the standards and internal procedures in force for that asset class.

PROPOSAL PATH · SUPERVISED AGENT PROPOSES CONF 0.82 SOLVER COMPUTES DETERMINISTIC PHYSICSNET GATE PASS / BLOCK RESPONSIBLE ENGINEER ACCEPTS OR REJECTS · SIGNED NO AUTONOMOUS PATH AGENT CANNOT REACH THE ASSET FIG 2 / SUPERVISED PROPOSAL PATH · NTS VIOLET = PROBABILISTIC · GREEN = VALIDATED · RED = BLOCKED EP-AGT-002
03 — WHAT IT MAY NOT DO

The limits are architectural, not configured

An agent that can be granted authority by changing a setting has authority. These constraints are properties of the system rather than options in it, so there is no permission an administrator can grant that lets an agent act on plant.

NEVERWrites a setpoint, issues a work order or actuates equipment
NEVEREdits the design basis or widens an operating envelope
NEVERPresents a prediction as a calculated result
NEVERReaches an engineer without a physics verdict attached

Blocked proposals are recorded too. A pattern of physics rejections is itself an engineering finding — usually about the asset, sometimes about the model.

04 — WHAT THE AGENT RETAINS

Context, and what it is allowed to conclude from it

RETAINED
USED FOR
STATUS OF OUTPUT
Design basis
Bounding every proposal to what the asset was built for
FACT
Operating history
Distinguishing normal drift from a developing fault
FACT
Past interventions
Not re-proposing what was tried and rejected, with the reason
FACT
Failure modes
Ranking causes against this asset's own FMEA
RANKED · CONFIDENCE
Remaining life estimate
Sequencing inspection against other assets in the fleet
PREDICTED · CONFIDENCE

The last column is the discipline: what the agent read is fact, what the agent concluded is not. Both travel with every proposal.

05 — SUPERVISION

Accountability does not move to the agent

Every asset has a responsible engineer, and every proposal is addressed to that person. Acceptance is a signed act with a timestamp; rejection is recorded with a reason and is fed back to the agent as context — never as training data for anyone else's model.

01

Addressed, not broadcast

A proposal goes to the engineer who holds the asset, not into a shared queue where ownership dissolves.

02

Rejection is information

The reason a competent engineer said no is the most valuable context the agent will ever receive, and it stays in the tenant.

03

The record survives the people

Who decided what, on what evidence, is retained for the asset's life — which is longer than most engineers' tenure on it.

06 — WHERE IT SITS

Layer 06 of seven — above the gate, below the engineer

Asset agents sit above PhysicsNET and ENGIN Solver and below the engineering decision record. They are the layer that composes; they are not the layer that concludes.

RECEIVES
Asset context from the knowledge graph Deterministic results from the solver Verdicts and blocks from PhysicsNET
EMITS
Proposals with sources and confidence Ranked failure modes, never single answers Its own rejection history, per asset
GUARANTEES
No autonomous action, by architecture Context confined to one asset and one tenant Every proposal traceable to its evidence

Point an agent at one asset you already argue about.

Design-partner evaluations start with a single asset that has a contested history — the case where accumulated context either earns its keep or does not.