CACV
Research
An experimental organisational model for autonomous commercial activity, with human authority over consequential resources and risk.
The research question
CACV, Coigach Autonomous Commercial Ventures, explores how specialised AI functions might undertake economically useful work within an explicit structure of human control.
The underlying Autonomous Economic System separates proposing and executing work from challenging it. Governing constraints remain outside the authority of the agents pursuing an objective.
Autonomy and authority
My central distinction is between operational autonomy and decision authority. A system might become more reliable at a bounded task without earning the right to change its risk limits or allocate consequential resources.
The design investigates how observed performance, independent challenge and human intervention could govern the amount of autonomy permitted. Poor performance or unresolved risk should reduce that permission rather than encourage a system to redefine success.
The economic test
Useful autonomy has to justify its oversight and computational cost. A result that appears successful may still be commercially unattractive if it requires constant human correction or creates risks that are difficult to reverse.
The work is at the architecture and experimentation stage. It is not evidence of a revenue-generating autonomous business or a proven capital-allocation system. Its contribution to my wider work is a way to examine delegated authority, incentives and review before granting systems broader operational freedom.