PAY

outcome impact outcome-driven Safety · maximize

Let AI move money → with a cap on every step

Metric type Safety, direction maximize. Every move an agent makes carries a per-transaction cap, a budget, and a human approval path, so control over money never means control without a limit.

BASELINE an agent pays with no limit set TARGET every move capped and recorded Safety ↑ maximize baseline no cap, no record target capped, logged, safe
Baselinean agent that spends without a cap is a risk you can't sleep on; one bad call and there's no record of why
MechanismPay puts a per-transaction cap, a budget, and a human-approval guardrail on both Collect and Spend, with an immutable audit trail of every move
TargetSafety maximized — money moves through AI with a limit on every step and a receipt for every dollar

Outcome chart · illustrative gauge, grounded in PAY's own claim · not a live metric