Spend limits
Per-transaction, daily, monthly, merchant, and category caps before the action is allowed.
AI agent spend controls
Stop agent overspend before money moves.
Featherlane AI adds a pre-action check between an AI agent and a money-moving tool. Your app sends the proposed amount, merchant, action, and context; the guard returns allow, cap, block, or hold before the charge fires.
Failure modes
The risk is not the answer. It is the action.
- A shopping agent retries a checkout link and creates a duplicate order.
- A travel agent books a non-refundable itinerary above the user cap.
- An AP agent pays an invoice with swapped bank details.
- A support agent issues a refund that policy says should be reviewed.
Runtime check
What the check sees
Policy effects
Return allow, cap, block, or hold-for-approval with a reason the app can show or log.
Action context
Check the amount, merchant, user intent, workflow step, and source evidence together.
Audit proof
Every decision leaves a trail
ok Trace id for each decision
ok Policy and cap that fired
ok Amount, merchant, and action metadata
ok Reviewer or approval context when a human steps in
Questions
What teams ask before they wire it in.
Where does Featherlane AI sit?
It sits before the tool or payment action fires. Your agent proposes an action, your app checks it with Featherlane AI, then applies the returned effect.
Does this replace payment rails?
No. Featherlane AI is a policy and audit layer that works with the rails your app already uses.
Ship the check
Put a effect between the agent and the action.
Bring a real checkout, refund, booking, or invoice flow. We will map the exact pre-action check and the audit record it should leave behind.