Adopting FINOS AIGF for Agentic AI in Financial Services
A practical guide to policy definition, architecture, deterministic orchestration, and observability for agentic AI an FSI organization can actually run in production — built around the FINOS AI Governance Framework (AIGF), CALM, and Fluxnova, and worked through end-to-end against a single KYC document-review use case.
First Edition — Draft. Contents subject to change ahead of first release.
- Introduction
- Chapter 1 — Governance-as-Code, in One Sentence
- Chapter 2 — Where AIGF Fits in the Pipeline
- Chapter 3 — The Determinism Problem, and Why Fluxnova Solves It
- Chapter 4 — What This Means for Engineering Teams
- Chapter 5 — What This Means for GRC Teams
- Chapter 6 — AIGF and AIUC-1: Differences and Complementarity
- Chapter 7 — A Step-by-Step Path Through the Pipeline
- Chapter 8 — Getting Teams Trained and Ready
- Chapter 9 — Codifying a Use Case
- Chapter 10 — Solution Architecture in Depth: From Reference Pattern to CALM Specification
- Chapter 11 — Fluxnova in Depth: Modeling the KYC Process
- Chapter 12 — Validating Controls in CI/CD: The KYC Build Gate
- Chapter 13 — Evals: Measuring What Determinism Doesn't Cover
- Chapter 14 — Observability in Depth: Evidence, Metrics, and the KYC Process in Production
- Chapter 15 — Tokenomics: Attribution, Cost, and Value
- Conclusion — Why This Matters More Than Another Framework
- References
Fifteen chapters and a running worked example — an agentic KYC document-review use case — carried from a use-case narrative through taxonomy classification, architecture, CALM specification, Fluxnova process modeling, CI/CD validation, eval-driven testing, production observability, and tokenomics. See the Introduction for the full shape of the argument, and the Conclusion for a chapter-by-chapter recap.
See CONTRIBUTING.md for this project's contribution process, DCO requirements, and governance policies.
Copyright © 2026 Fintech Open Source Foundation
This work is licensed under a Creative Commons Attribution 4.0 International License.
SPDX-License-Identifier: CC-BY-4.0.