2026-05-04T23:18:20.705Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:1:11807 2026-01-21T15:43:46Z 1

MULTI-AGENT AI FOR ADAPTIVE TREASURY AND CAPITAL OPTIMIZATION 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/1/11807 https://portal.finiz.singidunum.ac.rs/Media/files/2025/FINIZ-2025.pdf M. Mihajlović A. NJeguš Enterprises face growing challenges in managing liquidity and capital flows amid market volatility, regulatory complexity, and fragmented digital infrastructures. Recent industry research, including The Harmony Gap study by FIS and Oxford Economics, reveals that organizations incur nearly $100 million in annual losses due to inefficiencies such as cyber threats, fraud, and operational friction. This paper examines these challenges through the lens of business agility and introduces a multi-agent artificial intelligence (AI) framework for adaptive treasury and capital optimization. Using a modular, agent-based design, the system combines forecasting, liquidity management, FX (foreign-exchange) risk monitoring, and instrument recommendations to enhance visibility, resilience, and efficiency. By integrating data-driven forecasting with agent-based decision-making and rules-based safeguards, the framework aims to reduce idle balances, maximize yields, and improve compliance. We explore how multi-agent AI can transform treasury operations, enabling conferenceObject 78 87 10.15308/finiz-2025-78-87 Book of Proceedings Singidunum University International Scientfic Conference FINIZ 2025