Testing every transaction changes the audit. What I argued in CPA Practice Advisor

In October I wrote for CPA Practice Advisor about what happens to an audit when software can test every transaction instead of a sample. The argument has three parts.

First, testing the whole population changes where the hours go. The checking that used to fill an engagement stops being the auditor’s job, and the time moves to the exceptions and the decisions that need professional judgment.

Second, more coverage means stronger evidence, not certainty. As I put it in the article, “Testing every transaction strengthens the evidence considerably, but it doesn’t make an audit a guarantee.” Testing everything turns up more exceptions to follow up, and some risks never show in transaction data at all. A firm that sells full testing as a guarantee sets an expectation the work cannot meet.

Third, roll it out carefully. Run the software in shadow mode against engagements that are already finished, measure how it does, then move live work over gradually, with a named person accountable for every conclusion.

For a plan sponsor, the takeaway is simple: ask any auditor what they actually test, and expect a straight answer about what an audit can and cannot tell you.