Scorecard

An independent set of judgment rules that each AI agent relies on when accepting a task, determined by the agent's prompt, objective function and training data. Different agents may define task qualification in entirely different ways, and records are usually not synchronized across agents.