Home/Resources/Sample Action Receipt
The trust artifact

See what a run hands back.

This is the artifact Hubson gives you after a simulated or live run. It's designed for operators, managers, customers, and auditors — not developers reading raw logs. Below is a sample from a simulated refund-approval workflow.

Simulation receipt — produced entirely from sample data and mock app responses. No live apps or systems were touched.
Simulation receipt · no live apps or systems touched Live receipt · real actions completed under approval policy
hubson · Action Receipt · R-1284
Refund approval simulated · $2,840 · Stripe mock
requested by support teammate · via Zendesk #4821 · owner: Head of CS
✓ Simulation complete
Workflow
refund-approval
Run
R-1284 · simulation
SOP
finance.refunds.v3
Cost
0 run units · $0.0042
Outcome
refund approved
Run timeline
New request received00:00 · refund $2,840 · Zendesk #4821
Context pulled00:01 · plan Growth · Stripe charge in dispute window · billing error confirmed
SOP matched00:02 · finance.refunds.v3 · refunds over $2,000 require human approval
Stripe refund prepared, not sent00:03 · stripe.refund · approval-required · crm.delete_record restricted, not attempted
Paused for approval00:04 · over $2,000 threshold · held for owner
Approved by [email protected]00:08 · Head of CS · no edits — draft sent as written
Mock refund issued00:09 · mock stripe.refund · approved simulation
Customer reply drafted00:10 · zendesk.draft_reply · approved template
✓ Refund path approved · customer update drafted · follow-up: review dispute window in 7 days
Retained · exportable · reviewablesimulation · connect live apps to run for real

No live apps or systems were touched. This receipt was produced entirely from sample data and mock app responses.

What this proves

One receipt answers seven questions

  • Which SOP guided the workflow
  • Which apps/actions were reviewed or used
  • Why approval was required
  • Who approved it
  • What happened after approval
  • What it cost, and what follow-up is needed
Why the receipt is the point

Trust AI work because you can inspect it

Readable by anyone

No logs to decode — a manager, customer, or auditor can follow exactly what happened.

Model usage is visible

Every important run shows what it spent. Bring your own model accounts where supported, or see Hubson-managed model usage separately.

Approvals are on the record

Who approved, when, and whether they edited the action — captured in the receipt.

Reviewable later

Review the run later, inspect what happened, and repeat the workflow from the same configuration.

Produce your own

Run a simulation and get a receipt like this — free

Map your first workflow, simulate it on sample data, and read the receipt before connecting a single live app or system.