AEOM Scholarly Edition

For more than a century, organisational design has treated execution as something to define before work happens. Tasks, functions and processes each assumed that behaviour could be specified in advance. Agentic AI breaks that assumption: when intelligent actors can plan, reason and determine execution dynamically, prescribing every activity in advance stops being the organisation’s most valuable design act.

This Scholarly Edition sets out the full theoretical case for the Agentic Enterprise Operating Model (AEOM), building the formal apparatus behind it: Dynamic Agentic Workflow Notation (DAWN), The Box, Organisational Representation Theory (ORT), and Workflow Execution Memory (WEM).

The paper’s central construct, The Box, defines the minimum specification required for governed autonomous execution: Intent, Inputs, Outputs, Constraints, and Governance. Its most significant feature is not what it contains. It is what it deliberately omits: execution itself, left to emerge dynamically within governed boundaries.

Six contributions build the case:

DAWN and The Box: a minimal specification construct answering what intelligent actors need instead of a predefined process.
Specification-Centric Design: the proposition that specifications, not processes, become the primary and most stable organisational asset as execution becomes increasingly dynamic.
Human Sovereignty: a constitutional governance principle. Humans retain ultimate authority over purpose, governance, constraints, delegation and accountability; agents exercise authority only within delegated boundaries.
Organisational Representation Theory: a representational framework for Human, Agentic and Hybrid Actors, including Organisational Inheritance Theory (how authority, purpose and accountability propagate through organisational structure) and Organisational Polymorphism (how one role can be fulfilled by different actor types while preserving the same organisational contract).
Workflow Execution Memory: an evidence based organisational memory architecture turning execution into evidence, evidence into memory, memory into intelligence, and intelligence into learning.
A position on Agentic BPM: the paper maps DAWN’s constructs directly against the Agentic Business Process Management Research Manifesto (Calvanese et al., 2026) and sets out where AEOM extends that agenda, from process management toward enterprise design.

The paper works through a customer onboarding example from a regulated financial institution, contrasting a traditional process-centric workflow against the same capability expressed as an AEOM specification. It is also candid about where it runs against current market direction, treating Camunda’s ProcessOS launch as live evidence that the BPM vendor market is betting on making the process itself AI-native rather than replacing it with a specification layer, and framing AEOM’s claim as a contested prediction rather than a settled outcome.

Written by Michael White & Trevor Moore, Directors, IGX Solutions Ltd. Licensed CC BY-NC-ND 4.0. Share freely with attribution.