Agent Autonomy Lease, explained for beginners

Agent Autonomy Lease puts a person, not the AI, in charge of how much freedom and money an agent gets, and for how long.

What it is

Companies let AI agents work with different amounts of freedom: some may only read, some may write, some may spend money or take actions that cannot be undone. The question is: who decides that level, and can the agent change it?

Agent Autonomy Lease writes the level down as a lease: a digitally signed document from a person that says "this agent may act at level 3, spend up to $200 in total, until Friday". Every action is checked against the lease. The agent cannot sign itself a better lease. Small purchases are added up, so ten small ones cannot sneak past the total. An agent that keeps breaking the rules is automatically moved down a level until a person signs a new lease.

An everyday comparison

A rental car agreement: who may drive, how far, until when. The driver cannot rewrite the agreement, and the mileage adds up whether you drive in one trip or twenty.

How it works, in 4 steps

1. A person signs a lease: autonomy level, total budget, expiry.
2. Before each action, the agent's request is checked against the lease.
3. Spending is added up; sub-agents can only get a smaller lease than their parent.
4. Repeated violations lower the level automatically; only a new human-signed lease restores it.

Try it now

Open the five live demo tools → Every button sends a real request to the live product and shows the real answer.

See it on the homepage: Agent Autonomy Lease section →

Who it is for

Companies giving agents more freedom over time: operations, finance, IT automation, and platforms that run agents for customers.

FIRME AI CORPORATION · CAIN-42 is a Byzantine trust fabric. Numbers marked "our own check" were run by us, not by an independent party. Nothing on these pages is a certification.