KPMG brought nearly 1,000 summer interns in its audit and assurance track to its $450 million training campus outside Orlando, Florida, this year, replacing some of the routine bookkeeping drills long used to prepare new accountants with exercises built around judgment calls, teamwork under pressure and a fraud-investigation mystery interns had to solve as a team.
The redesign responds to a shift already underway inside the firm. KPMG has said it plans to hand off the most repetitive parts of an audit, the transaction-by-transaction checks that confirm payroll figures or revenue contracts match a company’s books, to artificial intelligence systems capable of running dozens of automated review tasks at once. A pilot version begins this year, with wider use expected in 2027 and a goal of shifting most routine checking to AI by 2029.
That timeline changes what an entry-level accountant does. Junior auditors have historically spent their early years tracing individual transactions, repetitive work that also taught them how a company’s finances fit together. KPMG has acknowledged that once AI absorbs that work, human auditors will spend more time reviewing what the software produces, gathering data and judging where financial statements carry the greatest risk of error, tasks built on comparing information and weighing competing explanations, judgment a checklist cannot supply.
The exercises, including a scavenger hunt and a team design project where the group loses a member partway through to mimic a co-worker pulled onto another assignment, are meant to build those instincts before interns sit with a client’s ledger. The approach also puts pressure on the offshore teams that currently perform much of the routine transaction testing, since that work is most exposed to automation. Whether KPMG retrains or redeploys those employees as the work disappears is a question the firm has not yet answered.
