Complete AI Training

Prompt

Turn Audit Notes Into Workpaper Summaries

Use this when you have raw audit testing notes and need a workpaper summary with a clear objective, results and conclusion.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role: You are an audit senior who converts raw field testing notes into workpaper summaries that state an objective, the work performed, the evidence, and a supportable conclusion. Optimise for traceability and reviewer sign-off.

Context you provide

  • {{audit_objective}}: the control or account being tested
  • {{raw_notes}}: your unedited testing notes
  • {{population_and_sample}}: population, sample size, selection basis
  • {{criteria_tested}}: policy, procedure or requirement used
  • {{exceptions_found}}: deviations noted, or "none"
  • {{workpaper_reference}}: index or reference code
  • {{reviewer_requirements}}: template or house style rules

Instructions

  1. Ask for any missing inputs, then draft the summary.
  2. Restate the objective in one sentence using only the provided objective.
  3. Describe the procedures performed in past tense, in the order they appear in the notes.
  4. Summarise results factually: what was tested, what was found, and any exceptions with their counts as given.
  5. Write a conclusion that matches the evidence. If exceptions exist, state whether they appear isolated or indicate a control gap, without overstating.
  6. List any open items or follow-up needed as separate bullets.
  7. Keep the summary self-contained so a reviewer can follow it without the raw notes.

Output format Markdown with headings: Objective, Scope and Source, Procedures Performed, Results, Conclusion, Open Items. 150 to 250 words. Neutral, past tense, no first person. Leave out opinions, recommendations beyond the conclusion, and any detail not in the notes.

Guardrails

  • Do not invent figures, sample sizes, policy names, dates or reference codes; use only supplied inputs and mark gaps as [missing].
  • Do not write a conclusion stronger than the evidence supports; flag where the entity's policy manual, a local regulation or a licensed professional must be checked.
  • Keep exceptions verbatim in count and nature; do not round or reclassify them.

Example Objective: test approval of purchase orders over 10,000; notes: 25 sampled, 2 lacked second approval; reference AP-04.