Turn a job-site voice note into an invoice draft
Capture completed work while it is fresh, then check the customer, rates and totals before billing.
What you’ll produce
An invoice draft with a transcript, matched customer, itemized work and missing-information list.
Inputs & connections
- A voice note or transcript describing completed work
- Approved customer records, rate card and invoice template
An approved recording or upload workflow plus an accounting connection. Begin with draft creation only; publishing invoices requires a separate approved action.
Your starting prompt
Replace the bracketed fields. This is our suggested prompt, not a quote from the source.
Turn [job note] into a draft invoice using [approved customer records, rates and template]. Extract the work date, customer, completed tasks, quantities and explicitly stated charges. Link each item to the transcript. Flag unclear customer matches and missing rates instead of guessing. Calculate totals from confirmed inputs and mark tax fields for review. Show the full draft for approval before creating records, sending an invoice or changing accounting data.
How to run it
- 01
Record the completed work and retain the original note.
- 02
Match the customer and service items against approved records.
- 03
Prepare the invoice draft and highlight missing or ambiguous fields.
- 04
Review the customer, quantities, rates and totals before approving any accounting write or delivery.
Review before you act
The owner checks that billed work was completed, customer details are correct and the approved pricing applies.
Limitations & success checks
- A voice transcription may confuse quantities, names or dates.
- The cited implementation is a vendor-reported customer example; this adaptation has not been reproduced and does not promise the same outcome.
Measure: Drafts accepted without correction, missing fields resolved and duplicate invoices prevented.
Sources & attribution
The sources below support the reported example or a tool capability. They do not validate our prompt or prove the outcome in your environment.
The 2024 customer story describes Wee Beastie Gardens using recorded requests, transcription and Zoho Books. Our draft-first checks are editorial additions; older model and pricing details are not current setup instructions.
Creator / publisher: Grace Miller / Zapier
October 6, 2026 product announcement establishes the current AI by Zapier naming and migration of agent capabilities into Zap steps.
Creator / publisher: Ruchira Kulkarni / Zapier
Editorial review: 2026-10-17. How we label evidence.