Applicable (Part 3): ATTRIBUTE 1 - Applicable

The Producer role is occupied by the agent itself, operating under human-defined parameters. And the Accepter role transforms: rather than reviewing a finished artifact at the end, the Accepter functions as an active supervisor — monitoring execution, setting intervention triggers, and retaining the authority to halt, redirect, or reject at any point in the process. This is less like quality inspection and more like air traffic control. The work is in motion. The human’s job is to keep it on course. Agentic AI Work is the frontier of the five types. Most organizations have not yet deployed agents at scale — but the ones that have are discovering quickly that their existing governance frameworks were not designed for work that acts before it is reviewed. WEEFAI is. The same Definer-Producer-Accepter structure applies; only the timing and character of each role shifts to match the nature of autonomous execution. These five work types — AI Assisted, AI Augmented, AI Generated, Non-AI Human Work Outcomes, and Agentic AI — cover the full spectrum of how knowledge work is now produced and delivered. A Planner sequencing a real workweek will encounter all five. The framework does not require separate systems for each. It requires a single system that can see all of them clearly and govern each appropriately. That is what Applicability means in practice: not that the framework bends to fit every context, but that it was built to hold all of them at once. There is, however, a sixth work type that inverts the usual human-AI relationship in a way none of the previous five address. AI Evaluated Output. In the previous five types, AI is always on the production side — assisting, augmenting, generating, or acting. In AI Evaluated Output, the human produces and AI reviews. A developer writes code that an AI security scanner evaluates against defined vulnerability criteria. A writer drafts content that an AI model checks against brand, compliance, or accuracy standards.