Traders Skeptical of Autonomous AI Trading Despite BNY Mellon, Microsoft Trial
Human traders remain essential to financial markets despite recent demonstrations showing AI agents can negotiate and execute trades.
While BNY Mellon and Microsoft successfully completed a trade between two AI agents earlier this month, industry professionals maintain that fully autonomous trading remains a distant prospect for the institutional sector.
The collaboration between BNY Mellon and Microsoft involved two AI agents agreeing on a price for US Treasury bills, according to reporting by Risk.net. These agents operated within strictly defined parameters, and human intervention remained a requirement to approve the final terms of the transaction. This structure highlights a persistent gap between technical capability and the practical requirements of market oversight.
Market participants point to the inherent risks of “rogue agents” and the continued value of human relationships as primary barriers to widespread adoption, as reported by KhanList. Financial institutions rely on complex, trust-based networks that software cannot currently replicate. Traders expressed skepticism regarding the speed at which these systems could move from experimental, constrained environments to high-volume production trading.
The current technological limitation centers on how trades are negotiated and cleared in a live environment. While the BNY Mellon and Microsoft trial proved that agents can agree on a price for specific instruments, the broader infrastructure of global markets demands a level of nuance that current large language models and autonomous agents struggle to provide.

The reliance on predetermined constraints suggests that the industry is still in a proof-of-concept phase.
Because the financial system is built on the assumption of human accountability, the transition to autonomous trading requires more than just technical success; it requires a fundamental rewrite of legal and clearinghouse protocols. Until these systemic issues are addressed, the role of the human trader as the final arbiter of value and risk is expected to persist.
The gap between the capability demonstrated in the BNY Mellon pilot and the reality of the trading floor is large. While the trial serves as a benchmark for what is possible, the industry continues to prioritize human-led oversight for the foreseeable future. Institutional players are expected to continue monitoring these developments, likely favoring incremental automation over a wholesale shift toward agentic trading.
The consensus among traders remains that while the underlying technology is progressing, the “long way off” assessment reflects a pragmatic view of the complexities inherent in global capital markets.
Founded in 1784 by a group that included Alexander Hamilton, the Bank of New York was the first company listed on the New York Stock Exchange in 1792. Through that lineage, BNY Mellon is regarded as one of the three oldest banks in the United States.
The modern corporation formed in July 2007 from the merger of the Bank of New York and Mellon Financial Corporation. As of September 2024, it held $52.1 trillion in assets under custody and administration, making it the world’s largest custodian bank and the first to surpass $50 trillion.
The industry will continue to observe how these experimental trades evolve in future fiscal quarters. As of November 20, 2026, there are no immediate plans to remove human oversight from the final approval stage of Treasury bill negotiations, ensuring that the human element remains the primary safeguard in market transactions.