Example processfinanceSOX

SOX ITGC audit automation software

An example of SOX ITGC audit as a governed process: agents do the 9 steps, and a person signs the gate before anything irreversible happens. Download it, import it into your workspace, and adapt it to your systems.

Download example · JSONImport into Minctrl →

minctrl.process/v1 · hash 841e6239581c…b5c1

The graph9 agent steps1 human gatesigns: Engagement partner (CPA)

SOX ITGC audit autopilot

SOX ITGC testing is the most repetitive part of assurance — access reviews, change evidence, control execution — yet only a licensed CPA may issue the ICFR opinion. Controls tested and exceptions severity-rated straight-through; the engagement partner signs the ICFR opinion before it is issued.

The governed flow · branches, parallel work & a human on the irreversible step
clean / low (deficiency only)significant / material weaknessrejected — re-test requiredopinion signedroll-forward to next period
Start
agent · low risk
Scope the in-scope ITGCs and pull access, change and operations evidence
parallel · fan-out / join
Fan out the three ITGC domains to test in parallel
agent · low risk
Test logical access ITGCs (provisioning, recerts, privileged access) and flag exceptions
agent · low risk
Test change-management ITGCs (approvals, segregation, migration evidence) and flag exceptions
agent · low risk
Test IT-operations ITGCs (jobs, backups, incident handling) and flag exceptions
parallel · fan-out / join
Join the three domains once all tests have run
agent · medium risk
Rate exception severity (deficiency / significant / material weakness); check independence and materiality
decision
Severity of exceptions?
agent · low risk
Auto-clear the control conclusion and draft a clean workpaper
agent · medium risk
Stage the deficiency package (severity rationale, exceptions, draft modified conclusion) for partner review
gate · human sign-off
A licensed CPA / engagement partner reviews the conclusion and signs the ICFR opinion — or sends it back for re-testsigns: CPA / engagement partner
agent · high riskirreversible
Issue the signed workpapers and the ICFR opinion
agent · low risk
Track remediation of deficiencies and roll forward to the next period
Done
low riskmedium riskhigh riskbranch / reworkgate · human sign-off
What it is modelled on

Concluding that a control is effective (or waiving a deficiency) is an audit judgment that can mislead a SOX filing, so an auditor or control owner signs off the control conclusion before it's recorded.

RuleWhat it requires
SOX Section 404 ITGCIT general controls over access, change management and operations must be designed and operating effectively, and management/auditors must be able to evidence that with tested controls.
Segregation of duties (SoD)The person who executes a change or transaction can't also be the one who approves it; toxic-access combinations are a reportable control weakness.
Change-management & access-review evidenceAccess recertifications and change approvals must be documented with who requested, who approved, and when — the evidence, not just the assertion.

A summary for orientation, not legal advice.

Use this example
01

Download the file

A minctrl.process/v1 document: the graph and the hash of the process it becomes.

02

Import it in Processes

Import into Minctrl opens this example in your workspace; one click stores it as your first version. Import JSON does the same with the file, and a file changed after download is refused.

03

Connect your systems

Every connector starts in shadow: steps run, nothing leaves Minctrl until you configure a connector and switch it live.

04

Adapt it

Change signers, time limits and steps on the canvas. Each save is a new version; a running case stays on the version it started with.

Questions
How does SOX ITGC audit automation gather evidence without becoming a black box?

The AI agent collects control evidence and tests access, change-management and segregation-of-duties controls, but every test carries the underlying evidence and a tamper-evident audit hash. Because runs are deterministic and replayable, a reviewer can re-run a test and get the same result and the same workpaper.

Can the AI conclude a control is effective on its own?

No. Concluding that a control passes — or waiving a deficiency — is an audit judgment that can mislead a SOX filing, so it parks at a control-conclusion gate for an auditor or control owner. The AI does the testing and exception-flagging; a person owns the conclusion.

How are segregation-of-duties conflicts handled?

The flow tests for toxic-access combinations where one person could both execute and approve, flags them as potential deficiencies, and routes them for human review with the supporting access data recorded — so SoD conflicts are evidenced rather than assumed away.

More examples