A Fleet, Not an Assistant (bartkolendowski.com)

A new paradigm in AI/ML workflow management is emerging, where agents operate autonomously, completing tasks without the need for human prompts. The author describes a system where various AI agents proactively execute tasks such as prospecting, data reconciliation, and even filing improvement proposals. This shift transforms the role of the human operator from that of a traditional assistant—responding to prompts—to a manager overseeing a fleet of autonomous agents, enabling a more efficient and scalable approach to task management. This development is significant for the AI/ML community as it highlights the growing autonomy of AI applications and the potential for significant productivity gains. The author underscores crucial management structures, including central inboxes for reports, gates to control irreversible actions, and a verification process to maintain trust in the outputs. As these autonomous agents take initiative, effective management becomes critical; building a system that allows for oversight and adjustment is necessary to harness the benefits of AI while ensuring accountability and alignment with business goals. This evolution in AI workflows suggests firms can achieve unprecedented leverage and operational efficiency, provided they adapt their management practices accordingly.