An agent earns its place by entering a task and producing an adopted result, recorded as evidence.
Agents are not a marketing feature; they are the network's execution amplifier. A network of 1,000 nodes, each able to direct five agents on structured work, holds output capacity equivalent to 5,000 execution units. That capacity feeds the same PoB and Carry settlement path as human work.
What agents amplify
Agents take on research, sourcing, outreach, follow-up, monitoring, and coordination. They do not chat; they enter scoped tasks and leave a full record. By scaling structured execution, they raise how much verifiable business a fixed set of nodes can advance.
An agent's value comes from entering a real outcome and being adopted, not from how large the model is.
Governance bounds
An agent operates inside strict bounds. These bounds are what let its output enter the record rather than stay informal.
How agent output settles
Agent work does not settle on a separate schedule. Its output joins the deliverable, passes through PoB reconciliation, and is attributed within the same Carry waterfall as human contribution.
Why this matters for the model
Agents lower the marginal cost of structured execution, which lets the network advance more outcomes without a proportional rise in cost. The gain is captured through verified outcomes, not through a token subsidy.
Treating an agent only as an efficiency tool understates its role. Inside the bounds above, it is part of the execution capacity that produces verified outcomes. It is also leverage that does not depend on a token.
An agent never settles on its own authority. No agent signs, transfers, or commits without human endorsement, and every action remains subject to reconciliation.