Glossary

What is a finance ontology?

The shared model of what your numbers mean: accounts, entities, periods, transactions and the rules that connect them. Why finance teams need one and how to start.

Eddie Ningombam Updated Nov 2026

A finance ontology is a shared, explicit model of the things a finance team reasons about (accounts, entities, periods, transactions, forecasts, variances and the business rules that connect them) and how they relate. It gives data from every system one meaning, so a number from the ledger, the CRM and a spreadsheet can be matched, traced and explained the same way every time.

Why finance needs an ontology

The same word means different things in different systems. "Revenue" in the billing system is not revenue in the general ledger, because the books carry accruals, deferrals and reversals. A "customer" in the CRM may be three records in billing. One system closes on fiscal periods, another reports by calendar month.

Today those differences are usually resolved by hand, in a spreadsheet, by the person who happens to know. A finance ontology writes the answers down once, as explicit definitions and rules, so every report, model and AI answer uses the same meaning.

What a finance ontology contains

ObjectWhat it representsExample
AccountA node in the chart of accounts, mapped to a standard classification6420 Software subscriptions
EntityA legal entity, business unit or cost center, with its parentUS subsidiary, cost center CC001
PeriodA fiscal period and how it maps to the calendarFY2026 Q3, July to September
EntryA single financial fact with its source referenceOne invoice line, posted to an account and period
Forecast or planA version of expected values with its assumptionsQ3 reforecast, version 2
VarianceThe difference between actual and plan, with its driversSoftware spend 12% over plan, driven by an early renewal
RuleBusiness logic, versioned with the date it took effectAllocation of shared IT cost by headcount
Lineage recordHow a number was produced, step by step, back to sourceSource file, transformations, rule versions used

Finance ontology vs. related ideas

ConceptWhat it coversWhat it leaves out
Chart of accountsThe list of accounts and their hierarchyEntities, periods, rules, plans and how systems map to each other
Data model or schemaHow tables and columns are storedWhat the data means in finance terms, and the rules applied to it
Semantic layerConsistent metric definitions for reportingVersioned business rules and lineage back to each transaction
Finance ontologyObjects, relationships, rules and their versions, in finance termsIt relies on source systems for the raw data

Why the rules must be versioned

Business rules change. Allocation methods get revised, a cost center is split, a new metric definition is agreed. If the rule is simply overwritten, last quarter's numbers silently change and comparisons stop being like for like.

Illustrative example

In November the team changes how shared IT cost is allocated, from revenue share to headcount. With a versioned ontology, November uses the new rule, October keeps the rule that actually produced the October numbers, and the reforecast comparison shows the rule change as its own line instead of hiding it inside the variance.

Why AI in finance needs an ontology

A language model working on raw tables has to guess what each column means. It will often guess well, and occasionally guess wrong with full confidence. When AI works on defined objects, such as this account, this entity, this period and this rule version, every answer can point back to the objects and rules it used. That is what makes an AI answer checkable.

This is the practical meaning of deterministic before probabilistic: define and reconcile the data with explicit rules first, then let AI reason on top.

How to start building one

  1. Pick one workflow, such as explaining variances on one expense category.
  2. List the objects it touches: accounts, cost centers, periods, vendors, plan versions.
  3. Write down the matching rules between systems, for example how a CRM account maps to billing customers and ledger entities.
  4. Write down the business rules that apply, with the date each took effect.
  5. Keep lineage: every reported number should record which source rows and rule versions produced it.

A finance ontology is the foundation layer of decision infrastructure.

Frequently asked questions

Is a finance ontology the same as a chart of accounts?

No. The chart of accounts is one part of it. An ontology also covers entities, periods, plans, variances, the business rules between them and how each source system maps in.

We already have a semantic layer in our warehouse. Do we need this?

A semantic layer standardizes metric definitions for reporting. A finance ontology adds versioned business rules and lineage to each transaction, which is what lets you explain and reproduce a number, not just display it consistently.

Who maintains a finance ontology?

Finance owns the definitions and rules, usually the controller or head of FP&A. Data or engineering teams maintain how source systems map into it.

E
Eddie Ningombam
Founder, PhrasIQ

Building InSightOS, the decision infrastructure layer for enterprise FP&A teams. Previously in finance operations and data infrastructure. Writing about decision latency, financial reasoning, and what it takes for FP&A to own the strategy conversation.

Get new articles in your inbox
FP&A strategy, variance analysis, and decision infrastructure. No noise.
✓ You're subscribed. First issue incoming.