https://cuicaihao.github.io/tags/decision-systems/Recent content in Decision Systems on C.CUI's LogHugoen-AUSun, 30 Aug 2026 00:00:00 +1000https://cuicaihao.github.io/posts/2026-08-26-ai-agents-in-the-enterprise-when-to-cede-control-over-next-steps/Sat, 29 Aug 2026 07:00:00 +1000https://cuicaihao.github.io/posts/2026-08-26-ai-agents-in-the-enterprise-when-to-cede-control-over-next-steps/When an enterprise adopts an AI agent, it is not merely granting access to tools. It is delegating the authority to choose subsequent actions dynamically as conditions change. An agent is preferable to conventional automation only when feedback-driven path selection creates real value—and when the enterprise can bound its execution space, inherit user permissions, trace state changes, and contain the blast radius of errors.