AI Limitations and Opportunities in Treasury (Valorean Technologies)

In this episode, Craig Jeffery and Arjun Krishnan discuss five key limitations of enterprise AI in treasury: data quality, pattern-based reasoning without true understanding, difficulty handling novel situations, variable outputs, and limited explainability. They also explore the opportunities and solutions behind these constraints, including AI-assisted data cleanup, anomaly detection, stronger controls, human oversight, and safer agent-based system design.

Enterprise AI for Treasury: A Guide to Agentic Implementation: https://amzn.to/4vTH8Fv

AI: The Automation Spectrum and Examples in Treasury (Valorean Technologies) (2026): https://strategictreasurer.com/420-ai-automation-spectrum-and-examples-in-treasury/

Valorean Technologies: https://valorean.ai/

Timestamps:

00:00 Introduction

00:34 AI limitations and opportunities

02:05 Data quality shapes AI quality

05:35 AI correlates but does not understand

09:29 Arjun Krishnan's background

10:18 AI and genuinely novel situations

12:37 Why AI outputs can vary