AI for Audit
Where AI sits in audit work and in the numbers being audited — scepticism, evidence and the file when a model is involved on either side. Orientation, not audit guidance.
Read it in the library →The Honest Map of AI in Audit
Where AI genuinely sits in the audit — the engagement walked stage by stage, the honest headline of full-population testing and its caveats, why a profession built on signatures and inspection is an unusually bad place for fluent guessing, and why most firm pilots die of governance rather than technology.
Using AI Without Surrendering the Judgement
How to use AI in audit fieldwork without giving up the thing the signature certifies — treating confident output as an unverified representation, tying every claim back to source documents, writing the file so a stranger can follow it, keeping client data inside the boundary, and testing the machine's reading the way you would test a junior's.
Auditing the Client's AI
The other half of the dual question — the client's models are already inside the numbers, so this module covers finding where they run, auditing estimates nobody can re-derive, testing controls whose logic moves during the year, and holding the management conversation that reveals which kind of client you have.
The File, the Firm and the Second Year
What holds the whole course together — the inspection that arrives years after the signature, the firm-level decision about which tools deserve reliance before any engagement leans on them, the second-year engagement where the client's model and your own tool have both quietly changed, and an honest account of which parts of the auditor's job compress and which parts stay stubbornly human.
Every lesson is free, with no sign-up. Reading happens in the library, where your progress is saved on your device.