OES is a canonical set of business objects, a Customer, an Order, a Product, an Invoice, and many more, defined once and understood by every connected system. It is the shared vocabulary that turns disconnected applications into one coherent business, and gives AI a safe, governed way to act on enterprise data.

What Is It

A canonical business language, published as a specification

The OneEnterprise Specification (OES) is a canonical model of the business objects every company already works with: Customer, Order, Product, Invoice, and many more. Each object has one agreed structure and meaning, independent of any single system.

Most platforms move data by mapping one system’s fields directly onto another’s, a fresh translation for every pair. OES replaces that with a single shared vocabulary. Every system translates to and from OES once, through a Business Connector that encodes that system’s real rules. The canonical object in the middle is the source of truth.

OES is a published specification that OneEnterprise governs. It is also protocol-agnostic: the same canonical objects work over MCP (for AI agents reaching enterprise systems) and over emerging agent-to-agent protocols. It is the semantic layer the integration world has been missing.

At a Glance
WHAT
Canonical business object specification
COVERS
Customer, Order, Product, Invoice, and more
GOVERNANCE
Published spec, OneEnterprise governed
PROTOCOLS
Works across MCP and agent-to-agent
ROLE
The shared language at the center of the platform

Customer   |   Order   |   Product   |   Invoice   |   Delivery   |   Payment   |   Vendor   |   Inventory

and many more canonical business objects

OES is one of three pillars

OES is the language. Business Connectors put it to work in each system. The OES-MCP Gateway opens it to AI.

THE LANGUAGE

OES

The canonical objects themselves. The agreed structure and meaning of every business record, defined once for the whole ecosystem.

THE TRANSLATORS

Business Connectors

Where OES meets each system. Connectors encode localization, tax, sequencing, and validation, translating between a system and the canonical object.

THE AI ACCESS LAYER

OES-MCP Gateway

How AI reaches the canonical objects. A governed gateway that lets agents read and act on enterprise data in business terms, not raw system fields.

What It Is For

A shared language for your systems, ready for AI

Integration has always been treated as a plumbing problem: connect system A to system B. The harder, more valuable problem is that those systems do not share a definition of what a Customer or an Order even is. OES exists to give them one.

Without a shared language
  • Every system pair needs its own bespoke field mapping
  • The same Customer means something different in each system
  • Adding a system multiplies the integrations to maintain
  • AI sees raw, system-specific fields with no business meaning
  • Mappings break with every upgrade or schema change
With OES
  • Each system translates to one canonical object, once
  • A Customer means the same thing everywhere it appears
  • Adding a system is one connector, not a web of mappings
  • AI reasons in business terms it can act on safely
  • The canonical model stays stable as systems evolve

This is why OES matters for the AI era. An agent that understands a canonical Order can work across every system that speaks OES. It configures pre-built, standardized connectors (setup time) rather than authoring fragile integration logic (design time), which is what keeps the AI story enterprise-safe.

When To Use It

Where OES does its work

OES is not a tool you switch on by itself. It is the layer underneath the Business Engine and the OES-MCP Gateway, and it earns its keep in these situations.

Connecting more than two systems

The moment a third system enters the picture, point-to-point mapping starts to multiply. OES keeps the cost flat: each system connects to the canonical language once.

Keeping shared records consistent

When the same Customer, Product, or Order must stay true across ERP, store, and CRM, OES provides the single definition the Business Engine synchronizes against.

Giving AI access to your systems

When you want agents to read and act on enterprise data, OES gives them a business vocabulary, so they operate on a canonical Order rather than guessing at raw fields.

Querying across systems at once

When a question spans several applications (“show me this customer’s full order history”), OES lets it be asked once, in canonical terms, instead of system by system.

OES powers the AI readiness ladder

The deeper an agent can go, the more it relies on a shared business language. OES is what lets it climb.

1

Access

AI reaches a system in that system’s own language. Live in production today.

2

Translate

AI works in business language through OES, not system-specific fields.

3

Operate

AI monitors and acts on processes in business terms across systems.

4

Orchestrate

AI configures and runs integration through natural language. The funded roadmap.

What Are The Benefits

What a shared language gives you that connections alone cannot

CONNECT ONCE

N, not N²

Every system connects to the canonical language once. The number of integrations grows in a straight line, not a web that multiplies with each new system.

BESPOKE MAPPINGS

No

Stop rebuilding a custom field map for every system pair. The canonical object is the contract both sides translate to.

A VOCABULARY AGENTS CAN ACT ON

AI-ready

AI works on canonical business objects, so it can reason and act safely across any system that speaks OES, governed every step of the way.

SURVIVES CHANGE

Stable

When a system is upgraded or replaced, the canonical model stays put. You update one connector, not every integration that touched it.

CONSISTENCY EVERYWHERE

One truth

A Customer means the same thing in every system, which is what makes real-time sync, clean reporting, and trustworthy AI possible.

A STANDARD, NOT A SILO

Open

OES is a published specification anyone can build against, designed to be contributed to the open standards governing AI access to enterprise systems.

Think of it the way banking thinks of SWIFT.

SWIFT did not invent banking. It standardized the messages banks already exchanged, so any bank could talk to any other. OES does the same for enterprise systems.

SWIFT for banking

A shared message standard meant a bank connected once and could transact with the whole network, instead of negotiating a custom format with every counterpart.

OES for business systems

A shared object standard means a system connects to OES once and can exchange with every other system, and with AI agents, in one canonical business language.

OES is the language.
The engines and the gateway put it to work.

The canonical objects come alive through the Business Engine that synchronizes them and the gateway that opens them to AI. Explore how the pieces fit, or read the manifesto behind the language gap.