Alexander Vityaz (ORCID 0009-0006-0489-7881) · Corezoid Inc., Dnipro, Ukraine Published: August 2026 · Version: v1 · License: CC BY 4.0
AI makes most companies dumber. Architecture makes them smarter.
Artificial intelligence deployed within a fragmented enterprise architecture amplifies fragmentation accumulated across software, integrations, processes, vendor dependencies, and organisational routines. This article defines the resulting burden as the fragmentation tax and connects it to the grounding problem: different functions and systems operate through incompatible projections of the same company.
Drawing on the Conant–Ashby Good Regulator Theorem, the article argues that effective regulation requires a coherent model of the enterprise. It defines a Smart Company as an organisation with a proactive, executable digital twin in which actors, processes, connections, states, and transitions between states are represented explicitly.
This model gives the company two central properties: plasticity, which makes organisational structure available for deliberate change, and a Company Brain, which unifies memory, computation, routing, simulation, and the actions of people, software, and AI within a single operational loop.
API visibility is proposed as a practical diagnostic of whether the company possesses a coherent model of itself. The article concludes that AI becomes useful when embedded in an executable architecture that provides holistic operational context.
Keywords: Smart Company; executable digital twin; digital twin of the organisation; enterprise architecture; organisational cybernetics; Good Regulator Theorem; fragmentation tax; grounding; Company Brain; Actor Graphs; artificial intelligence; API architecture; metaprogramming
This is an essay rather than a formal paper: it states the architectural argument in plain terms and carries no theorems of its own, drawing instead on the results proved elsewhere in this corpus.
| File | Description |
|---|---|
| paper.pdf | Canonical PDF (identical to the Zenodo deposit) |
| paper.md | Readable markdown version |
Vityaz, A. (2026). How to Become a Smart Company. Zenodo. https://doi.org/10.5281/zenodo.21792643
@misc{vityaz2026smartcompany,
author = {Vityaz, Alexander},
title = {How to Become a Smart Company},
year = {2026},
month = aug,
publisher = {Zenodo},
doi = {10.5281/zenodo.21792643},
url = {https://doi.org/10.5281/zenodo.21792643}
}- Builds on On the Necessity of Noise Suppression for Minimal Good Regulators — supplies the formal necessity behind "building a model is not a charming optional extra, it is a mathematical necessity" (reference [1]).
- Builds on Actor Graphs — the executable model whose explicit actors, states and transitions make organisational structure available for metaprogramming (reference [2], §4.1; the record's first revision cited Active Transaction Graphs here, replaced by the Actor Graphs paper in the August 19 revision).
- Builds on A Phase Model of Enterprise Evolution — the architectural and economic consequences of fragmentation (reference [3], §2).
- Builds on Company Brain — the unified memory and environment for thinking that this essay names as the company's second property (reference [4], §4.2).
- Builds on The Computable Boundary of the Firm — the digital twin as the transactional architecture the argument depends on (reference [5]).
- Builds on Management Debt — Part I — the accumulated cost of fragmentation, here named the fragmentation tax (reference [6], §2).
- Cited by The Compact Company — §3.11 draws on this essay for the fragmentation tax and for the distinction between local functional amplification by AI tools and the architectural integrity of the company (reference [61]).
- Version of record: https://doi.org/10.5281/zenodo.21792643
See CHANGELOG.md.