How should an audit firm start using AI?

Question

Accepted Answer

Start with one real audit workflow where the problem is clear, the evidence is available and the result can be reviewed. Do not begin with a firm-wide promise to “transform audit”.

ITGC assessments and process understanding are strong starting points because both contain substantial manual work and create visible outputs. A firm can define its methodology, test the workflow on live engagements, compare the results with its existing approach and gather feedback from preparers, reviewers and clients.

A sensible first rollout has five parts:

- Choose a workflow with a clear owner and measurable pain.
- Configure the tool to the firm's methodology and evidence requirements.
- Pilot it across a representative set of engagements, not only the easiest files.
- Review quality, exceptions, time saved and user behaviour.
- Expand only once the firm understands what good use looks like.

Senior sponsorship matters, but day-to-day adoption is won by making the work easier for audit teams. The first use case should create practical value quickly while teaching the firm how to govern AI in production.

Firms can begin with a focused Platformed workflow—typically ITGC assessment or process understanding—configure it to their methodology and test it on representative live engagements before expanding.