Miquel Noguer I Alonso (Artificial Intelligence Finance Institute) and Harry Mendell (Federal Reserve Bank New York) have posted “Extending the FAIR Framework: Financial Agentic Systems” on SSRN. Here is the abstract:
This paper extends our Finance-Aware Implementation and Remediation (FAIR) framework by addressing a critical gap in current agentic system design: the temporal challenges inherent in real-world financial operations. While existing demonstrations focus on static tasks like research reports, financial markets require systems that adapt to continuously changing conditions-from equity trading execution to supply chain procurement. We provide practical implementation guidelines for financial institutions adopting Large Language Models (LLMs) and autonomous agentic systems, with particular emphasis on temporal safeguards that prevent stale data dependencies, execution drift, and cascading temporal failures. The enhanced framework introduces conditional execution protocols, dynamic re-validation mechanisms, and graceful degradation strategies that mirror successful approaches in autonomous vehicles and algorithmic trading systems.