Paper: arXiv 2609.27632
Authors: Walter Kurz, Reinhard Magg
Abstract
We present a compliance-first architecture for AI in regulated finance that treats regulation as an orientation layer rather than a deterministic ruleset. A matrix of regulatory intent and exposure provides a compact classification handle, which a governed policy compiler then maps into concrete prohibitions, obligations and runtime budgets. Prohibitions constrain feasibility and block externalisation, while obligations extend tasks with artefacts that must meet explicit admissibility criteria. Committee activation remains policy-driven and proportionate, preserving efficiency while ensuring supervisory oversight. Evidence, decisions and reason codes are bound to a permissioned DAG with deterministic timestamping, enabling replay, provenance checks and clear attribution of failure. Clause-level legal indexing with effective dates and capability-based agent routing ensure portability across DACH and the wider EU. The result is assurance by construction: compliance is embedded in execution and verifiable by auditors without sacrificing proportionality or transparency.
Complexity vs Empirical Score
- Math Complexity: 3.0/10
- Empirical Rigor: 2.0/10
- Quadrant: Philosophers — conceptual discussion, lower math and empirical scores
Why this score: This paper presents a novel conceptual framework for compliant AI in finance, focusing on architectural design rather than mathematical models or empirical validation. While the ideas are innovative and well-articulated, there’s a lack of concrete mathematical derivations or empirical backtesting to support the proposed architecture’s effectiveness. The paper outlines a system design but doesn’t provide details on implementation or data for replication.
Research Flowchart
flowchart TD
A["Research Goal: Compliant AI in Regulated Finance"] --> B{"Compliance-first Architecture Design"};
B --> C["Inputs: Regulatory Intent & Exposure Matrix, Legal Indexes"];
C --> D{"Computational Processes: Governed Policy Compiler, Permissioned DAG"};
D --> E["Outcomes: Assurance by Construction, Verifiable Compliance"];
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a compliance‑first AI architecture for regulated finance that treats regulation as an orientation layer, using a regulatory intent and exposure matrix to generate a compact handle v=(t,e) that is compiled by a governed policy compiler into concrete prohibitions, obligations and ru
Source sentence
We present a compliance-first architecture for AI in regulated finance that treats regulation as an orientation layer rather than a deterministic ruleset.
Assumes 1
- Only mild regularity is assumed, namely that dominant disqualifiers in P⊥ remain policy invariant for a given task state, so p(π,τ)=p(π′,τ) for all policies π,π′. execution
Source sentence
We assume only mild regularity, namely that dominant disqualifiers in P ⊥remain policy invariant for a given task state, so that p(π, τ) = p(π′, τ) holds for all π, π′ ∈Π; in practice these include conditions such as age thresholds, sanctions hits, or bans tied to specific product tiers.
Data
No dataset named in the text.
Costs
Not applicable
Finds 1
- The architecture turns compliance from after‑the‑fact attestation into assurance‑by‑construction, with reasoning, decisions and evidence cryptographically bound to a permissioned DAG for replay, provenance checks and precise failure attribution. theoretical
Source sentence
The proposal turns compliance from after-the-fact attestation into assurance-by-construction, where reasoning, decisions and evidence are bound cryptographically so reviewers can check not only outcomes but also whether the agent’s understanding and follow-through were correct.
Checked by others
No other paper in the archive has been found to support or contradict these findings. Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (52)
- Generic AI-DLT Enterprise System: Architecture and methodology for scalable domain adaptation from a unified core framework (2025)
- Financial and Operational Impacts of Regulatory Compliance on the Austrian Securities Industry (2025) — link
- Financial integration and structure in the euro area (2024) — link
- BIS international banking statistics and global liquidity indicators at end-September 2024 (2025) — link
- European Union: Market access and equivalence (2025) — link
- Regulation (EU) No 575/2013 on prudential requirements for credit institutions and investment firms ( CRR ) (2013) — link
- Directive 2013/36/EU on access to the activity of credit institutions and the prudential supervision of credit institutions ( CRD ) (2013) — link
- Directive 2014/65/EU on markets in financial instruments ( MiFID~II ) (2014) — link
and 44 more
- Regulation (EU) No 600/2014 on markets in financial instruments ( MiFIR ) (2014) — link
- Directive 2009/138/EC on the taking-up and pursuit of the business of insurance and reinsurance ( Solvency~II ) (2009) — link
- Directive 2009/65/EC on undertakings for collective investment in transferable securities ( UCITS ) (2009) — link
- Directive 2011/61/EU on Alternative Investment Fund Managers ( AIFMD ) (2011) — link
- Regulation (EU) No 648/2012 on OTC derivatives, central counterparties and trade repositories ( EMIR ) (2012) — link
- Regulation (EU) No 909/2014 on improving securities settlement in the European Union and on central securities depositories ( CSDR ) (2014) — link
- Directive (EU) 2015/2366 on payment services in the internal market ( PSD2 ) (2015) — link
- Directive 2009/110/EC on the taking up, pursuit and prudential supervision of the business of electronic money institutions ( EMD~II ) (2009) — link
- Directive (EU) 2015/849 on the prevention of the use of the financial system for the purposes of money laundering or terrorist financing (4 AMLD ) (2015) — link
- Directive (EU) 2018/843 amending Directive (EU) 2015/849 (5 AMLD ) (2018) — link
- Regulation (EU) 2023/1114 on markets in crypto-assets ( MiCA ) (2023) — link
- Regulation (EU) 2022/2554 on digital operational resilience for the financial sector ( DORA ) (2022) — link
- Regulation (EU) 2016/679 on the protection of natural persons with regard to the processing of personal data ( GDPR ) (2016) — link
- MiFID~II, Article 16: Organisational requirements (Interactive Single Rulebook) (2023) — link
- Commission Delegated Regulation (EU) 2017/565 supplementing Directive 2014/65/EU (2017) — link
- Commission Implementing Regulation (EU) 2021/451 laying down ITS for supervisory reporting under the CRR (2021) — link
- MiFID~II, Article 34: Freedom to provide investment services and activities (passporting) (2023) — link
- PSD2, Article 28: Application to exercise the right of establishment and freedom to provide services (2024) — link
- Gesetz über das Kreditwesen ( KWG ) (2025) — link
- Wertpapierhandelsgesetz ( WpHG ) (2025) — link
- Securities Trading Act ( WpHG )—overview (2019) — link
- Bankwesengesetz ( BWG ) (2025) — link
- Securities Supervision Act 2018 ( WAG 2018) [convenience translation] (2024) — link
- Financial market supervision in Austria (integrated model) (2025) — link
- Federal Act on Financial Services ( FinSA ) (2024) — link
- Federal Act on Financial Institutions ( FinIA ) (2024) — link
- FINMA’s legal basis (Financial Market Supervision Act and sectoral laws) (2025) — link
- Equivalence of non-EU financial frameworks: overview and list of decisions (2025) — link
- Trends in European investment funds: Fact Book 2025 (2025) — link
- EIB Investment Survey 2024: Country overview—Austria (2025) — link
- Proportionality in banking regulation: a cross-country comparison (2018) — link
- Evaluation of the effects of financial regulatory reforms on SME financing (2019) — link
- A novel multi-agent hierarchical deep reinforcement learning framework for portfolio optimization (2025)
- Multi-agent based deep reinforcement learning framework for trading strategy and portfolio management (2024)
- Recent advances in reinforcement learning in finance (2021) — link
- Systematic review on reinforcement learning in the field of FinTech — → 2305.07466
- Audit logs management and security: a survey (2021) — link
- Ensuring data integrity and auditability in BPM systems: A blockchain-based approach (2025)
- Is permissioned blockchain the key to support the external audit process? (2022) — link
- Transforming auditing through AI and blockchain (2025) — link
- Digital Operational Resilience Act ( DORA ) (2025)
- The EU’s Digital Operational Resilience Act ( DORA ) (2024)
- Regulatory Technical Standards on ICT services supporting critical or important functions (2024)
- Claude Code (2024)
1 of 52 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.