Industry · Industrial and energy

Find the last cheap maintenance window, not just the failure.

Condition monitoring tells you a reading has passed a limit. AnantState shows how long an asset has been heading there and when the last inexpensive window to fix it is.

Book a working session Compare the other industries

In plain termsKnow the last cheap moment to maintain a machine, not just when it breaks.

1 · What are the things you manage?

The things a maintenance planner decides about.

Machine
Production line
Asset
Site
Work order
Shipment
Energy meter or feed

2 · What changes over time, and why?

State	Baseline drivers	Unexpected drivers
Condition and degradation	Duty cycle, runtime hours, ambient conditions	Bearing wear, contamination, misalignment
Yield and quality	Recipe, feedstock, throughput	Drift in a process parameter, feedstock variation
Unplanned-stop risk	Maintenance interval, age	Cascading failure, operator variance
Energy posture	Load profile, tariff, weather	Equipment inefficiency, leakage

This is the domain that most resembles classical condition monitoring, and where the approach differs most clearly: the physics-informed baseline is genuinely available and the residual is genuinely small. A model asked to correct a known degradation curve is doing a much easier job than one asked to invent it.

3 · What goes wrong today?

In the buyer’s own words.

“We have vibration sensors and a threshold. We find out at the threshold.”
Condition monitoring says when a limit has been passed, not how long the asset has been heading there or which window is the last cheap one.
“Maintenance is scheduled by calendar because we cannot justify scheduling by condition.”
Moving to condition-based maintenance means pricing the intervention against the cost of waiting, and nobody has those numbers.
“A stoppage is explained afterwards, in a meeting.”
The machinery to replay the decision that led to the stoppage does not exist.

4 · What changes with AnantState?

Asset state as a forward position, so the last cheap maintenance window is visible, not just the impending failure.
Priced interventions. Inspect, derate, schedule maintenance, expedite a part, each with its cost, compared against running to failure.
Counterfactual replay of a stoppage, which turns a post-mortem argument into a measurement.
State across an estate, so “which of 400 assets gets the next maintenance crew?” is answered in one queue ordered by value at stake.
Yield and energy as state, using the same machinery, because a drifting process and a wearing bearing are the same problem shape.

5 · What can you see in the demo?

A synthetic industrial world of lines and assets: condition state, degradation trajectories, priced maintenance interventions, and the cost-of-waiting comparison across an estate.

Reference worlds are fictional and labeled.

Integration reality check

Plant data is frequently network-segmented. The read-mostly, no-write-back posture on the integrations page is why this is deployable in an operational-technology environment at all.
Latency needs vary from milliseconds (interlocks) to hours (maintenance planning). We target the planning and prioritization layer, and we say so plainly.

The last cheap window

Degradation shows as the gap between the expected path and the observed one, long before a limit is passed.

A threshold tells you when it has already happened. The gap between where an entity should be and where it is exists much earlier, while the cost is still small.

Price the maintenance window

Bring one asset class and its maintenance cost. We will show the cost of waiting against the cost of acting now.

Book a working session See the OgMart reference world