Accounting technology / Glossary / AI in auditing

What is AI in auditing?

Definition

AI in auditing is the use of machine learning and language models in audit work: scoring every transaction in a ledger for risk, matching evidence documents to the items tested, reading contracts and minutes, and drafting workpapers. The auditor stays responsible for the opinion and must be able to explain and document what each tool did.

Also calledAI auditAI-assisted audit

Compare the softwareBest audit software →All 8 audit vendors →

US standards: PCAOB, fiscal years from 15 Dec 2025
UK guidance: FRC, June 2025
Category: Audit

How AI is used in audits

Audit firms used rules-based data analytics for a decade before generative AI. The current tools add three things. Full-population testing: MindBridge scores every transaction for risk rather than a sample. Evidence matching: DataSnipper matches invoices, bank statements and confirmations to workpapers inside spreadsheets. Workflow and drafting: Fieldguide, Modus and Caseware use AI to prepare requests, test and tie out schedules and draft documentation. Regulators have responded. In the US, the PCAOB amended its standards on technology-assisted analysis for audits of fiscal years beginning on or after 15 December 2025; in the UK, the FRC published its first guidance on AI in audit in June 2025.

Common questions

Is AI being used in auditing?+

Yes. The large firms run their own AI platforms, and mid-size and smaller firms buy tools for risk scoring, document matching and workpaper drafting. Use is greatest in journal entry testing, evidence matching and summarising documents.

Will AI replace auditors?+

No. Auditing standards make the auditor responsible for the evidence and the opinion. AI changes the work: more of the population is tested and less time goes on matching and ticking, while judgement and review stay with people.

What do auditors have to document when they use AI?+

What the tool did, the data it used, how its reliability was checked and how its output was reviewed. The FRC's 2025 guidance sets out proportionate documentation for AI tools, and PCAOB standards require auditors to evaluate the reliability of electronic information used as evidence.