AI Native. AI Augmented. Partner Directed. Innately Audit.

Our agents read the records you have given us and draft the work the plan requires: an understanding of your business, an assessment of where the financial statements could be materially wrong, a plan of which balances need auditing and how, and the tests that plan calls for. Work is often re-performed, and other agents reconcile it, before a person reviews and then relies on any of it.

No human is perfect and no AI is perfect. Every section is drafted and then checked by multiple agents — often the same item more than once — so that one agent's blind spot is caught by another. Human faults are caught. Knowledge is never lost between person or team, as our agents never forget. Ultimate adoption and approval is always a human activity.

It does not decide. Every test, judgement, recommendation and conclusion becomes part of the audit file only when a named auditor adopts it as their own work. The software is a force multiplier. It allows us to do more, faster, more efficiently and more intelligently.

It does not sign. Every working paper records who prepared it, who reviewed it, and how far that review went. Those are people. Purely helped by an army of specialised agents with access to many documents.

It does not run unapproved. Every tool is certified by the firm for a specific purpose, and runs on your audit at the direction of our audit team.

Your records are never used to train a model, by us or by the vendors we use. Inference runs inside the UK and EU — the credentials our application holds are technically prevented from routing your data anywhere else, not merely configured not to.

The Financial Reporting Council expects firms using AI in audit to be able to explain what their tools do, when they are appropriate to use, and how the firm satisfied itself they work. We think our clients are entitled to the short version of the same answer.

If you want the long version, ask. Your auditor can share our tool certification documentation.