ClairAudit runs the procedures of an ISA-aligned audit with AI agents, and holds every finding to the same test a reviewer would: which clause, which evidence, who signed.
Finding F-07 · Medium
Covenant headroom narrows under the downside scenario within twelve months of the reporting date. Evaluate disclosure of the related events and conditions.
Standards corpus
ISA 570 (Revised), paragraph 16
Where events or conditions are identified that may cast significant doubt on the entity's ability to continue as a going concern, the auditor performs additional procedures to determine whether a material uncertainty exists.
Evidence E-12
Lender covenant compliance certificate, Q2
sha256 3f9a…c41e
An AI output is only as useful as the record behind it. ClairAudit keeps that record for every finding, so an engagement quality reviewer can follow the reasoning back to its source.
How a finding is madeLedgers, confirmations and documents arrive through connectors or the client request portal, each stored with a content hash.
A procedure agent works the evidence and proposes findings with severity, amount and rationale.
Each finding must link to a clause in the standards corpus. Uncited findings are sent back; if still uncited they are marked unvalidated.
A second model scores the finding against its evidence and clause. Low scores route to human review instead of passing silently.
Low score: routed to human review
Managers and partners approve, reject or override at gates. Nothing is signed on their behalf.
Every action appends to a hash-chained, write-once trail, verified daily, with Merkle roots anchored to public timestamping.
Modules follow the audit cycle, from acceptance to the report, and each is labelled with the standard that governs it.
See the capabilities
Natural-language questions are turned into a structured query plan, checked against an allow-listed schema, and executed under row-level security. The model writes the plan; it never writes SQL against your data.
How queries are guardedWhich open findings rated high are older than 30 days, by engagement?
plan table: findings
filter: status = open, severity = high
filter: createdAt < now - 30d
group: engagementId
scope tenant (RLS)| Engagement | Open high findings |
|---|---|
| Northgate Logistics FY25 | 3 |
| Harbour Foods FY25 | 1 |
A demo covers your audit methodology, the standards you report under, and how agents, citations and review gates would fit your team. Pricing is discussed on the call.