Plain guide

What is AI usage proof?

AI usage proof is a verifiable record of what an AI did for your business: which actions ran, who approved them, and what changed. Riley records it honestly, so you can show what your AI actually did.

Start free

AI usage proof is a verifiable record of what an AI did for your business: which actions ran, who approved them, and what changed. It turns a vague claim like "the AI handled it" into evidence a client or a manager can check.

Riley is the AI coworker that proves its work. It records approval decisions, signs receipts of completed work, and keeps a per-key access record, so the answer to "what did your AI do?" is a document, not a memory.

Who this is for

For anyone who has to account for the AI

If AI touches work that other people rely on, sooner or later you have to show what it did.

Agencies

Agencies running AI for clients

Show each client what the AI did on their account, and that a person on your team approved the decisions that mattered.

Local business

Local businesses and owners

Keep a clear record of what got answered, fixed, and posted, so you always know what happened while you were busy.

Ecommerce and SaaS

Ecommerce and SaaS teams

Give a manager a usage picture of AI-assisted work without asking anyone to reconstruct it from memory.

Consultants

Consultants

Hand over a verifiable record at the end of an engagement instead of a summary the client has to take on trust.

The problem

"The AI handled it" is not an answer

When AI does work on a real account, the value is easy to lose. The action runs, the moment passes, and by the time someone asks what happened, the only record is what a person remembers. That is fine until a client questions the bill, a manager asks for an audit, or a mistake needs tracing.

The honest version is harder than it sounds. It is not enough to log that "something ran." Real usage proof has to show what was proposed, who approved it, what actually changed, and how you know. Anything less is a claim dressed up as a record.

How Riley helps

Proof falls out of the work

You do not assemble usage proof by hand. Riley records it as a byproduct of the same loop it runs to do the work.

  • Connect your tools and pick the outcomes you want Riley to handle.
  • Riley runs the work you approve, and the approval is recorded.
  • It verifies each change landed, so the record reflects reality.
  • It signs a receipt you can share, backed by an access record per key.
Recorded today

What Riley records right now

Honest statuses only. Here is what is live today and what is still rolling out.

Approval decisions

Live

Every action that writes outside your workspace passes through the approvals queue. What was proposed, who approved or rejected it, and when, is recorded.

Signed receipts

Live

Completed work is assembled into signed receipts with a public verify link. That is a record you can hand to a client and they can check it themselves.

Per-key access, with last used

Live

Each AI surface connects with its own per-workspace key. See which keys exist, when each was last used, and revoke any of them from Settings.

Per-surface usage logs

Coming soon

A per-surface breakdown of which AI client asked for what, over time, in one view. This is rolling out and is labeled coming soon everywhere until it ships.

A concrete example

An agency proves a month of work

From "trust us" to a link the client can open

  1. Over the month, Riley drafts and posts review replies and listing fixes across a client's locations, each one approved by the agency.
  2. Every approval is recorded, showing which team member said yes and when.
  3. Each completed action is verified on the live source and rolled into a signed weekly receipt.
  4. At the review call, the agency sends a public verify link instead of a slide, and the client checks it directly.
  5. If anything could not be confirmed, the receipt says so plainly, which is what makes the rest worth trusting.
The proof it creates

Three answers, on the record

Put together, Riley's records answer the three questions any usage audit asks. Approval history answers who signed off. Signed receipts answer what actually happened. Per-key access answers which surface did it and when. Each one is recorded as the work happens, and each carries an honest note about anything that could not be confirmed.

Common questions

AI usage proof FAQ

What does Riley record today?

Riley records three things right now: approval decisions, meaning what was proposed and who approved or rejected it; signed receipts of completed work with a public verify link; and per-key access, so you can see which AI surface connected and when each key was last used. A per-surface usage log is rolling out and is labeled coming soon until it ships.

Why does AI usage proof matter for agencies?

When you run AI on behalf of clients, someone eventually asks what it did and whether a person stood behind the decisions that mattered. Approval history answers who signed off, signed receipts answer what actually happened, and per-key access answers which surface did it and when. That is the difference between saying AI helped and proving what it did.

Is AI usage proof the same as a proof receipt?

They are related. A proof receipt is a signed weekly record of the work Riley completed. AI usage proof is the wider picture: the receipts plus approval decisions and per-key access records. Together they let you show both what was done and that it was done under your control.

Can I share the proof without giving up access?

Yes. Every signed receipt carries a public verify link that anyone can open to confirm it is real, without an account or a login. You choose what to share and with whom.

Show what your AI actually did.

Put your work on one loop and let the proof record itself. Start free, no card to begin.