Methodology paper
DIG: The Declarative-Imperative Graph
A universal modelling methodology for the whole enterprise for the AI era.
Abstract
An enterprise is not only its processes. It is its products, customers, people, commitments, risks, regulations, contracts, controls, and the typed web of relationships that binds these. Each of those domains is represented somewhere in the enterprise's software estate today. None of them represents the enterprise as a whole. Each is a view. There is no single model behind the views.
DIG, the Declarative-Imperative Graph, is a universal modelling methodology that represents the enterprise as a whole in a single graph, with every domain modelled using the same primitives. It rests on three moves: it unifies declarative structure (rules, regulations, obligations) with imperative structure (events, observations, transactions); it lifts the entities this knowledge is about into high-dimensional, state-bearing, lifecycle-governed citizens of the graph; and it is itself a graph-native data layer rather than a consumer of one.
From this design follow three properties an organisation needs before it can put an AI agent anywhere near a decision someone will later have to defend: traceability, reliability, and determinism.
Cite this paper
@unpublished{jonsson2026dig,
author = {Jónsson, Ragnar Láki},
title = {DIG: The Declarative-Imperative Graph},
subtitle = {A universal modelling methodology for the whole enterprise for the AI era},
year = {2026},
month = {may},
url = {https://ahead.is/dig}
}