Your client is paying for AI work they cannot see. What record proves what the AI actually did?

What is an audit trail for AI agency work?

It is not a folder of screenshots. It is not a chat export. An audit trail for agency work is a call by call record tied to a deliverable, where each entry can be checked by someone outside the agency. Each AI tool call leaves a receipt: the tool that ran, the time, the exact saved result, and the sha-256 digest of that result.

Scope, stated once: an agent receipt is a record of one ai agent tool call that anyone can check: it keeps the exact saved bytes and their sha-256 digest, and it proves the result was not changed since it was saved, not that it was right.

That scope is what makes the trail useful in a client relationship. The agency is not asking for trust. It is handing over a record the client can test, one digest at a time, at zambo.dev/verify, free, with no account.

Why agencies need this now

Agency work used to leave its own trail: drafts, files, time entries, version history. AI work often leaves almost nothing a client can inspect. The model ran, text appeared, and the steps in between live in a chat window the client never saw.

• Billable work needs backing. When hours turn into AI calls, the client still wants to know what was done on their account.
• Approvals need a fixed record. Sign-off means little if the deliverable can quietly change after the client approved it.
• Disputes need facts, not memory. When a client says the work was never delivered, or was different, both sides need the same record to point at.
• Auditors need more than a story. A compliance review asks what ran and what it returned. A narrative answer is not an audit trail.

What a complete trail looks like, deliverable by deliverable

Start with the deliverable and work down. A report, a campaign plan, a research brief: each one should resolve into the calls that produced it.

1. The call list. Every tool call behind the deliverable, in order, with times.
2. The saved results. What each call actually returned, kept as saved bytes, not summaries of summaries.
3. A receipt per call. Each result carries its digest, so any entry can be checked on its own.
4. A handover chain. HANDOVER, Zambo flagship product, links the calls and the deliverable into one chain the client can follow and sign off on.
5. A dispute pack when needed. If a payment or a scope question turns into a dispute, the same trail exports as the pack both sides work from.

Zambo runs this as one MCP endpoint at https://zambo.dev/mcp, 153 tools total, 35 on the default surface, with a receipt on every call in the AER-1 format, specified by the IETF Internet-Draft draft-zambo-aer1-26.

Worked example: one receipt from a live work chain

From the public stream on October 25, 2026: a workflow_close call saved at 2026-10-25T18:28:07Z, receipt at zambo.dev/run/26c308be-922e-4cd3-a171-053b7b4ec0fa. A workflow_close is the kind of call that closes out a chain of work, which is exactly where an agency trail matters: the moment work becomes a deliverable.

The verification endpoint returned verified for this receipt. Run the same check yourself at zambo.dev/verify. Picture a month of client work built this way: every deliverable opens into its calls, every call opens into its receipt, and any receipt can be checked by the client in seconds. That is the audit trail, and it needs no meeting to explain.

How to put this in place at an agency

Add Zambo to the AI your team already uses at zambo.dev/add. Calls start issuing receipts right away, 20 free calls per day to start. A Day Pass is $0.99 USDC on Base for 24 hours of full access. The Zambo Pass is $29 for 30 days, one payment, with an annual option at $290.

Then make it a habit, not a rescue move: attach the handover chain to each deliverable, collect sign-off against that chain, and keep the dispute pack ready before anyone asks. Agencies that do this stop arguing about what the AI did. The record answers first.