AnantState
Ontology

You name the things you manage. The platform does the math.

Before the platform can model your operation it needs your vocabulary: what a store, a lane or an account is, what is measured about it, how things connect, what actions exist and what each costs. You write that down once, you own it, and you can change it without waiting for a software release.

In plain termsA shared dictionary of your business, written by you. (Engineers call this an ontology.)

Five declarations

What the ontology contains.

  • 1 · Entities

    The nouns the operation is managed through.

    A store, a lane, an account, a machine, a case, a pathway. Each has an identity and a lifetime.

  • 2 · State dimensions

    What each entity carries, its bounds, and which dimension is primary.

    The primary dimension is the one that carries business value and therefore drives ranking and pricing.

  • 3 · Relationships

    How entities connect.

    A region contains stores; a supplier feeds lines; an account shares a device with another. Relationships make impact analysis and cascade reasoning possible.

  • 4 · Dynamics

    How state moves, in the business’s own terms.

    Schedule, seasonality, throughput, wear, decay, contract terms, known events.

  • 5 · Interventions and policies

    What actions exist, what each costs, and the limits it may be applied within.

    Including what each is permitted to affect, and what may be automatic.

A new world is a definition, not a release. The business declares what exists, how it moves and what acting costs; the platform does the arithmetic.

Four things the ontology unlocks

Governed like the asset it is

If the vocabulary is not governed, nothing built on it is either.

PropertyWhy
VersionedA change to a state dimension changes what every downstream conclusion means.
ReviewableAdding an intervention creates a spend path; it needs a second pair of eyes.
AttributableEvery change records who made it and when.
ScopedTenant and domain boundaries are respected; one business unit’s vocabulary does not leak into another’s.
ReversibleA change can be rolled back, because the previous version is a real artifact.

A platform that lets anyone redefine “primary dimension” in production has no meaningful audit story, however good its models are.

Building an ontology is a workshop, not a project

  1. Name the entities

    Usually the nouns people already argue about.

  2. Name the state

    The three to six numbers that decide whether someone acts.

  3. Pick the primary dimension

    The one that carries money. This is contentious, and worth the argument.

  4. Declare the dynamics

    How state moves when nothing unusual happens. Domain knowledge, not data science, does the work here.

  5. List the interventions and price them

    Usually the most revealing hour of the engagement, because most organizations have never priced their own actions.

  6. Declare the horizon

    How far ahead must this be right?

Most domains reach a first, useful ontology in a workshop. That is intentional: the barrier to a useful twin should be domain knowledge, which you have, not data science, which you may not.

Run the six-step workshop on one domain

Bring the nouns your team argues about weekly. We will leave the session with a first ontology and a forward position for each entity.