All products

Prepare audit workpapers with AI

Plan a control test, document how samples were selected and give the reviewer a clear trail from each result to the evidence examined.

To prepare audit workpapers with AI, describe a control in terms someone can test. For example, the finance team expects a second person to review a bank reconciliation every month. With audit-support, the working paper can identify which months belong in the population, what approval should look like and where to record its date.

A sound design and a missing approval are different questions. The first asks whether the control could address the risk. The second asks whether the business can show that it operated during the period. Keep them separate so a reviewer can follow the evidence and decide what additional work is necessary.

What it does

  • Describe the risk and purpose of a control
  • Define the population and period for testing
  • Compare random, systematic and targeted sample selection
  • Record exceptions beside the evidence inspected
  • Prepare review questions and follow-up work

Install and use

  1. Install the Finance plugin in a compatible agent and provide an authorised control description and population. The agent consults this reference when a relevant finance request arises.
  2. Ask “Draft a workpaper for monthly bank reconciliation approval. Leave space for evidence, review date and the reason each sample was chosen.”
  3. Agree the selection method with the auditor, inspect the original evidence and have a qualified reviewer assess any exception.

FAQ

Source

Made by Anthropic. Original skill: audit-support, published on GitHub under the Apache-2.0 licence.

View on GitHub