Enterprise business, general

Why this page is horizontal

Every other domain page in this set reuses an industry vocabulary. This one does not have one, because the questions it answers are not industry questions: what is an entity, what is it called, who owns the definition, who decides, what has to exist before work moves on.

The reuse layer here is thinner and should be described honestly. schema.org is a shared vocabulary for describing things, widely embedded in web markup and useful as a lowest common denominator. gist is an upper ontology for business concepts maintained by a consultancy. BORO and REA are modelling methods from practice — BORO for business objects, REA for resources, events and agents in accounting. None of these is a consensus standard of the kind ISO 25964 is, and a team adopting one is adopting a house style, not a norm. Where the vocabulary work itself needs a published rulebase, that comes from library and information science.

The layer most organizations skip: roles and handoffs

The April 2026 preprint cited below separates its model into three parts: a role layer, a domain layer and an interaction layer. The domain layer is where most projects stop. The interaction layer is where the preprint puts handoff patterns, approval chains and escalation paths — a directed edge from one role to another, with a trigger condition, the artifacts that must exist, and whether approval is required.

Set out plainly, a handoff is small: a design-to-development handoff, triggered on design complete, requiring the interface designs and specifications, with the product manager as the escalation path. Held as a typed edge, it is something an agent can follow. Held as a paragraph in a process document, it is not.

This is the same argument as the rest of the site, applied one layer up. An agent that can compute the right answer but cannot tell who authorizes it has not been given enough context to act, and will route the decision to whoever it likes.

The coupling ladder, as a checklist

The same preprint proposes six levels of coupling between a formal model and a model's reasoning. The levels are not validated, and they are useful for one purpose: making a team state which level it is actually claiming.

Two things are worth lifting out. The first is that L1 to L2 is the step from giving an agent better text to giving it fewer wrong options, and it is cheaper than it sounds. The second is that the ladder's top rungs need the constraint layer this site describes under SKOS and its neighbours: without shape validation, L4 is a claim rather than a mechanism.

The playbook

Reuse

schema.org for broad interchange, gist or BORO for upper-level business concepts as a house style, REA where accounting semantics matter, and ISO 25964 and SKOS for the vocabulary method behind all of them.

Scope

One role's decision surface, not the org chart. Pick the role whose decisions are most often routed to the wrong place.

Competency questions

Who approves this class of decision? What artifact must exist before the handoff? At what threshold does it escalate, and to whom? Which role sees this data, and which must not?

Smallest artifact

A role and handoff model: roles as entities, handoffs as typed edges carrying trigger, required artifacts and approval. Add the vocabulary of decision types so that “approve” means one thing.

Validation

Replay real decisions from the past and check whether the model routes them to the approver who actually signed. Where it disagrees, the model is wrong or the documented process was.

Characteristic failure

Role blindness: the same correct data framed for the wrong reader. A finance answer handed to an engineer and an engineering answer handed to a finance controller are both wrong in the same way, and neither looks like an error.

Limits of this page

• The architecture and the ladder both come from a single April 2026 preprint, self-reported by its platform's authors. The role-and-handoff argument does not depend on its numbers, and the numbers should not be quoted as settled.
• The reuse layer here is the weakest in the set. Where library science, finance and healthcare have published norms to align to, this page has practice artifacts and judgment, which makes it the most expensive of the ten per unit of scope.
• Nothing here addresses permissions and access control, which most organizations solve in an identity system rather than in a knowledge model, often with consequences.