Skip to content

Operating ontology

One ontology, from data to decision.

The same objects and links sit under everything I build, so finance, AI development, and delivery work as one system instead of three tools bolted together.

01

From ontology to delivery

A shared model of finance, program, and control objects feeds the agents built on top of it, which in turn feed the surface where work actually gets reviewed and shipped.

A three-stage flow. A shared ontology of finance, program, and control objects feeds AI development, which feeds AI delivery. Strategic finance and governance run across all three stages.

Strategic finance · Governance & model risk

The connective tissue across every stage

01

Ontology

Objects, links, and the data model

Entities & relationships

Capital, plans, programs, and controls as linked objects.

Strategic finance model

Allocation, forecasts, and metrics defined once.

Operating & program data

Owners, milestones, risks, and variance.

02

AI Development

Building agents on the model

Agentic workflows

Agents that act on the ontology, not loose text.

Model risk & evaluations

Test cases and checks before anything ships.

Evidence capture

Prompts, sources, and assumptions, logged.

03

AI Delivery

Operating it in the business

Control plane

One surface for the work and its state.

Decision loops

Review, approve, and report on a cadence.

Audit trail

Every decision reconstructable after the fact.

Controls · Auditability · Speed — held to the same bar as everything else

02

The same system, in layers

Read bottom to top: the ontology is the foundation, finance and program operations sit on it, AI development builds on that, and delivery runs at the top. Governance holds across all of it.

A four-layer stack. From the bottom: an ontology and data model, then strategic finance and program operations, then AI development, then AI delivery. Governance, evidence, and controls run across every layer.
L4

AI Delivery

Control plane, decision loops, and executive reporting.

L3

AI Development

Agentic workflows, evaluations, model risk, and evidence capture.

L2

Strategic Finance & Program Operations

Capital allocation, metric definitions, and program execution.

L1

Ontology & Data Model

The objects, links, and actions the whole system shares.