pith. sign in

arxiv: 2606.07119 · v1 · pith:MPPH3JUBnew · submitted 2026-06-05 · 💻 cs.ET · cs.AI· cs.MA

The Three-Ring Architecture: Governing Agents in the Era of On-Platform Organisations

Sergio Alvarez-Telena , Marta Diez-Fernandez This is my paper

Pith reviewed 2026-06-27 20:08 UTC · model grok-4.3

classification 💻 cs.ET cs.AIcs.MA
keywords three-ring architectureagentic AIgovernance layerstrategies-based agentsLLM riskorganisational operating systemfederation layerenterprise compliance
0
0 comments X

The pith

A strategies-based federation layer is required as the operating system for governing agentic enterprises.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that organisations acquiring agentic AI lack the infrastructure to govern it, reproducing earlier AI deployment failures at a projected 95 percent rate. It formalises the Three-Ring Architecture in which Ring 2, built on strategies-based agents, performs the functions of resource abstraction, process coordination, permission enforcement, and stable intelligence compounding at the organisational level. Ring 2 is presented as necessary rather than optional because its deterministic framework allows traceable consequences, enforceable permissions, and recoverable deviations, in contrast to the non-deterministic risks of Ring 3 LLM-based agents. The architecture has been deployed over a decade in financial services, government, procurement, and compliance.

Core claim

Ring 2 constitutes the operating system of the agentic enterprise by performing at the organisational level what a computing OS performs at the device level: resource abstraction, process coordination, permission enforcement, and a stable platform for compounding intelligence. Strategies-based agents operate within a deterministic framework whose consequences are traceable, permissions enforceable, and deviations recoverable. LLM-based agents introduce a categorically distinct non-deterministic risk that propagates through complex organisational systems without retrospective traceability. Ring 2 is therefore a necessary condition of control and compliance.

What carries the argument

The Three-Ring Architecture, with Ring 2 as the M2 federation layer of strategies-based agents that functions as the organisational operating system separating deterministic and non-deterministic risk profiles.

If this is right

  • Improvements in LLM capability increase the structural requirement for Ring 2 governance.
  • The architecture supplies the control layer needed for on-platform organisations to compound intelligence safely.
  • Deployment experience across financial services, government, procurement, and compliance demonstrates practical applicability.
  • Absence of the federation layer produces decentralised intelligence without enforceable compliance.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Governance investment should focus on building the deterministic Ring 2 layer before scaling Ring 3 capabilities.
  • The model may apply to regulating agent interactions across multiple independent organisations.
  • Enterprises could treat strategies-based agents as the primary control surface rather than an auxiliary tool.

Load-bearing premise

Strategies-based agents operate inside a deterministic framework whose consequences remain traceable and recoverable while LLM-based agents produce non-deterministic deviations that cannot be traced retrospectively.

What would settle it

A documented case in which LLM-based agents receive retrospective traceability and recovery mechanisms that match the deterministic properties attributed to strategies-based agents, or a case in which strategies-based agents fail to coordinate processes at enterprise scale.

Figures

Figures reproduced from arXiv: 2606.07119 by Marta Diez-Fernandez, Sergio Alvarez-Telena.

Figure 1
Figure 1. Figure 1: The Three-Ring Architecture. Ring 2 (solid boundary) is the OS: deterministic, governing, architecturally definitive.Ring 1 (inner dashed) is contained within Ring 2 and progressively absorbed. Ring 3 (outer dashed) is a loosely coupled satellite: probabilistic, replaceable, governed by but not contained within Ring 2. AI can enter at any ring. disrupting the underlying systems. Over time, users cease to i… view at source ↗
read the original abstract

The current phase of enterprise AI deployment faces a structural failure: organisations are acquiring agentic capability without the infrastructure to govern it. The result is expected to reproduce the error of the first wave of AI deployment: decentralised intelligence without a federation layer leading to a 95% project failure rate. This paper formalises the Three-Ring Architecture as the governing infrastructure of the on-platform organisation. Ring 1 is the existing production architecture; Ring 2 is the M2 federation layer built on strategies-based agentic AI; Ring 3 is the LLM-based frontier intelligence layer. Ring 2 constitutes, in the technically exact sense, the operating system of the agentic enterprise - performing at the organisational level what a computing OS performs at the device level: resource abstraction, process coordination, permission enforcement, and a stable platform for compounding intelligence. A central contribution is the formal distinction between Ring 2 and Ring 3 risk profiles. Strategies-based agents operate within a deterministic framework: their consequences are traceable, their permissions enforceable, their deviations recoverable. LLM-based agents introduce a categorically distinct risk: a non-deterministic actor whose deviations propagate through complex organisational systems without retrospective traceability. Ring 2 is not a useful addition - it is a necessary condition of control and compliance. A further consequence: every improvement in LLM capability is a structural tailwind for this architecture. More capable non-deterministic actors produce larger consequences when they deviate. The governance requirement scales with capability. The architecture has been validated across a decade of deployment in financial services, government, procurement, and compliance among other sectors.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes the Three-Ring Architecture as the governing infrastructure for on-platform organizations deploying agentic AI. Ring 1 is the existing production architecture; Ring 2 is the M2 federation layer built on strategies-based agentic AI, which the paper claims functions exactly as the operating system of the agentic enterprise by performing resource abstraction, process coordination, permission enforcement, and providing a stable platform; Ring 3 is the LLM-based frontier intelligence layer. The central claims are a formal distinction in risk profiles—strategies-based agents are deterministic with traceable, enforceable, and recoverable consequences, while LLM agents produce non-deterministic deviations without retrospective traceability—and that Ring 2 is a necessary condition of control and compliance rather than an optional addition. The architecture is asserted to have been validated across a decade of deployment in financial services, government, procurement, and compliance.

Significance. If the risk distinction and necessity claims were supported by formal models, definitions, and empirical evidence, the paper could provide a useful conceptual framework for addressing governance gaps in enterprise agentic AI deployment, particularly by framing improvements in LLM capability as increasing the structural need for a federation layer. The identification of decentralized intelligence without a federation layer as a recurring failure mode is a reasonable observation, but the absence of derivations or data in the current manuscript limits any assessment of practical significance.

major comments (3)
  1. [Abstract] Abstract: The central claim that strategies-based agents operate within a deterministic framework whose consequences are traceable, permissions enforceable, and deviations recoverable is asserted without a definition of 'strategies-based agentic AI', a formal model of how determinism is enforced at the organizational level, or any example demonstrating traceability and recovery. This distinction is load-bearing for the necessity of Ring 2.
  2. [Abstract] Abstract: The assertion that Ring 2 'constitutes, in the technically exact sense, the operating system of the agentic enterprise' performing resource abstraction, process coordination, permission enforcement, and a stable platform is presented as a formal contribution but rests on definitional mapping rather than derivation from benchmarks, external data, or mechanisms, rendering the necessity argument circular.
  3. [Abstract] Abstract: The claim that the architecture 'has been validated across a decade of deployment in financial services, government, procurement, and compliance' is stated without any metrics, methodology, case details, or counter-examples, which is required to support the empirical grounding of the central claims.
minor comments (2)
  1. [Abstract] The term 'M2 federation layer' is introduced without expansion or prior definition.
  2. The manuscript would benefit from a table or diagram explicitly mapping the functions of Ring 2 to OS properties to clarify the analogy.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the detailed and constructive comments. We respond to each major comment below, indicating where we will revise the manuscript to address the concerns while preserving the conceptual focus of the work.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that strategies-based agents operate within a deterministic framework whose consequences are traceable, permissions enforceable, and deviations recoverable is asserted without a definition of 'strategies-based agentic AI', a formal model of how determinism is enforced at the organizational level, or any example demonstrating traceability and recovery. This distinction is load-bearing for the necessity of Ring 2.

    Authors: We agree that a precise definition would strengthen clarity. In revision we will add an explicit definition of strategies-based agentic AI as systems whose behavior is governed by explicit, human-specified strategies that produce deterministic execution traces. We will also include a concise organizational example showing traceability and recovery. A full formal model of organizational determinism lies outside the scope of this architecture paper, which prioritizes the risk-profile distinction over mathematical derivation. revision: yes

  2. Referee: [Abstract] Abstract: The assertion that Ring 2 'constitutes, in the technically exact sense, the operating system of the agentic enterprise' performing resource abstraction, process coordination, permission enforcement, and a stable platform is presented as a formal contribution but rests on definitional mapping rather than derivation from benchmarks, external data, or mechanisms, rendering the necessity argument circular.

    Authors: The operating-system characterization is offered as a functional analogy to communicate the layer's role in resource abstraction, coordination, and permission enforcement at organizational scale. The necessity claim is not derived from the analogy but from the preceding risk-profile distinction: absent a deterministic federation layer, non-deterministic deviations from Ring 3 become untraceable. We will revise the text to separate the analogy from the logical argument and to make the inference steps explicit, thereby removing any appearance of circularity. revision: partial

  3. Referee: [Abstract] Abstract: The claim that the architecture 'has been validated across a decade of deployment in financial services, government, procurement, and compliance' is stated without any metrics, methodology, case details, or counter-examples, which is required to support the empirical grounding of the central claims.

    Authors: The statement reflects the authors' accumulated experience implementing comparable governance structures in the cited sectors. Because the manuscript is a conceptual architecture paper rather than an empirical study, no metrics or case details were supplied. We will revise the abstract to indicate that the architecture is informed by such deployments, removing any implication of formal empirical validation. revision: yes

Circularity Check

1 steps flagged

Ring 2 necessity and OS equivalence asserted definitionally from risk distinction

specific steps
  1. self definitional [Abstract]
    "Ring 2 constitutes, in the technically exact sense, the operating system of the agentic enterprise - performing at the organisational level what a computing OS performs at the device level [...] Ring 2 is not a useful addition - it is a necessary condition of control and compliance. A central contribution is the formal distinction between Ring 2 and Ring 3 risk profiles."

    The OS mapping and the deterministic/non-deterministic distinction are introduced as definitional properties of the rings; the necessity of Ring 2 is then derived from those same properties. The distinction is presented as the 'formal' contribution yet consists of the assertion itself, with no separate model or data showing how determinism is enforced or traceability achieved at organisational scale.

full rationale

The paper's central claim that Ring 2 is a necessary OS layer for control reduces directly to its own definitional framing of the three rings and the asserted deterministic vs. non-deterministic distinction between agent types. No equations, external benchmarks, or independent derivation establish the traceability/enforceability properties or the OS analogy; these are stated as the architecture itself. The 'formal distinction' and 'technically exact sense' are therefore self-referential. The decade-long validation claim supplies no metrics or methods. This produces partial circularity (score 6) without requiring self-citation chains.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

The central claim rests on the introduction of three conceptual layers and the domain assumption of widespread governance failure, with no free parameters or independently evidenced entities.

axioms (1)
  • domain assumption Organisations are acquiring agentic capability without the infrastructure to govern it, reproducing the first wave of AI deployment errors at a 95% project failure rate.
    Stated directly in the opening of the abstract as the motivating structural failure.
invented entities (2)
  • Three-Ring Architecture no independent evidence
    purpose: Governing infrastructure for on-platform organisations consisting of production, federation, and frontier layers.
    Newly formalized framework introduced to solve the stated governance gap.
  • M2 federation layer (Ring 2) no independent evidence
    purpose: Strategies-based agentic AI layer that functions as the operating system of the agentic enterprise.
    Core invented component whose functions are defined by analogy to computing OS.

pith-pipeline@v0.9.1-grok · 5825 in / 1529 out tokens · 29540 ms · 2026-06-27T20:08:55.783467+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

13 extracted references

  1. [1]

    and Díez-Fernández, M

    Alvarez-Teleña, S. and Díez-Fernández, M. (2025). Advances in Agentic AI: Back to the Future https://arxiv.org/abs/2512.24856. arXiv

    arXiv 2025

  2. [2]

    and Díez-Fernández, M

    Alvarez-Teleña, S. and Díez-Fernández, M. (2025). Advances in Geostrategy: Extreme-Efficient Nations https://ssrn.com/abstract=5136657. SSRN

    2025

  3. [3]

    and Díez-Fernández, M

    Alvarez-Teleña, S. and Díez-Fernández, M. (2026). Advances in People Management: Embrace Rotation https://ssrn.com/abstract=6653538. SSRN

    2026

  4. [4]

    and Díez-Fernández, M

    Alvarez-Teleña, S. and Díez-Fernández, M. (2022). Data MAPs: On-Platform Organisations https://ssrn.com/abstract=4232955. SSRN

    2022

  5. [5]

    Edward N., L. (1963). Deterministic nonperiodic flow https://journals.ametsoc.org/view/journals/atsc/20/2/1520-0469_1963_020_0130_dnf_2_0_co_2.xml. Journal of the Atmospheric Sciences

    1963

  6. [6]

    Forrester, J.W. (1961). Industrial Dynamics http://www.laprospective.fr/dyn/francais/memoire/autres_textes_de_la_prospective/autres_ouvrages_numerises/industrial-dynamics-forrester-1961.pdf. MIT Press

    1961

  7. [7]

    MAy, R.M. (1976). Simple mathematical models with very complicated dynamics https://www.nature.com/articles/261459a0. Nature

    1976

  8. [8]

    Álvarez-Teleña S., Systematic Trading: Calibration Advances through Machine

  9. [9]

    Álvarez-Teleña S., Díez-Fernández M., Data MAPs: On-Platform

  10. [10]

    Álvarez-Teleña S., Díez-Fernández M., Advances in Portfolio Management:

  11. [11]

    Álvarez-Teleña S., Díez-Fernández M., Advances in

  12. [12]

    Álvarez-Teleña S., Díez-Fernández M., Advances in Cognitive Warfare:

  13. [13]

    & Funk, R.J., Papers and patents are becoming less

    Park, M., Leahey, E. & Funk, R.J., Papers and patents are becoming less