SOX 404 Compliance
Scoping, control design, testing and remediation for Section 404, including the first-year programmes that decide whether year two is manageable.
SOX 404 support
Access, change and operations controls tested by someone who can read both an access listing and a general ledger.
Access, change and operations controls over the systems that produce financial data, which is what IT general controls are. Everything above them depends on them: automated controls and system-generated reports are only as reliable as the ITGCs underneath.
The work needs someone who can read both an access listing and a general ledger. An accountant who asks for the listing without being able to interpret it, and a technologist who understands the system but cannot tie a finding to a financial statement assertion, both produce reports that miss the point.
An IT audit and ITGC consultant is usually brought in because a financial statement auditor has issued a request list, and nobody internally is confident about which items on it are the company's job to produce or what a good answer to them would look like.
A weakness in these controls is pervasive rather than isolated, and that is the part that surprises people. One access review that was never performed does not stay a small finding about one review. It undermines every automated control and every system-generated report that rests on the same access model, which is how it turns into a material weakness.
Who can get into the system, what they can do once inside, and whether that has been checked lately. In practice: provisioning tied to an approval, de-provisioning tied to a leaver process that actually runs, privileged and administrative accounts inventoried and justified, periodic access reviews performed by someone who understands what the entitlements mean, authentication standards, and (persistently the most common finding) generic and shared accounts that cannot be attributed to an individual.
The access review is where most programmes fail, and it fails in a specific way. The review is performed by a manager who receives a list of usernames and role codes, cannot interpret them, and approves the list. The control exists on paper and provides nothing.
How code and configuration reach production. Changes authorised before they are made, tested before release, approved by someone other than the developer, and migrated by someone who cannot also write the code. Emergency changes are the recurring weak point, legitimate in principle, but frequently the route by which normal change control is bypassed with retrospective documentation that was never actually reviewed.
Job scheduling and monitoring of failures, backup execution and (importantly) restore testing, incident management, and the ability to demonstrate that data processed completely and accurately. A backup regime nobody has tested a restore from is a plan, not a control.
Where the financially significant application is a SaaS product, the traditional ITGC questions change shape. You do not control the change management process; the vendor does. You do control who has access, how your tenant is configured, what integrations move data in and out, and whether you obtain and read the vendor's SOC 1 or SOC 2 report.
That last item is where most companies stop short. Obtaining the report is treated as the control. It is not. The report contains complementary user entity controls, things the vendor explicitly states you must do for the vendor's controls to be effective. Those are your responsibility, they are rarely read, and an auditor who reads them will ask how each is satisfied.
Infrastructure-as-a-service brings its own scope: IAM policy, network configuration, logging and monitoring, and infrastructure-as-code change control. That overlaps heavily with the cloud security compliance work, and where both are in scope they are tested once rather than twice.
Twenty years auditing banks, insurers and mortgage companies produced familiarity with a regulatory environment that is more prescriptive than most: FFIEC IT examination expectations, GLBA safeguards obligations, third-party and vendor management programmes that examiners test in detail, business continuity and resilience requirements, and the specific scrutiny applied to core banking, loan origination and claims systems.
Institutions in Los Angeles County and Ventura County preparing for an examination, or responding to findings from one, are a recurring client type. The value is knowing what an examiner will actually look at; which is narrower and more specific than the published guidance suggests.
The biotech and medical device concentration along the 101 corridor through Thousand Oaks, Newbury Park and Camarillo generates a distinct variant of this work. Systems holding regulated records fall under FDA 21 CFR Part 11, which imposes requirements (validation, audit trails, electronic signature controls, record retention and system access) that overlap substantially with ITGC but are assessed against a different standard and by a different kind of inspector.
Companies frequently run these as two disconnected programmes: a validation effort owned by quality and an ITGC effort owned by finance, testing many of the same controls twice against different documentation. Mapping them once removes a meaningful amount of duplicated work.
A control matrix mapping each ITGC to the financial statement assertions and application controls it supports, so that the significance of a failure is visible rather than assumed. Testing workpapers with evidence retained to a standard an external auditor can rely on. A findings register distinguishing design gaps from operating failures, because they require different remediation. And a remediation plan that accounts for what the IT team can realistically deliver alongside its existing commitments.
Identify which applications and infrastructure are financially significant. Scoping every system in the company is the most common and most expensive error in this work.
Walkthroughs with IT and process owners covering how access is granted, how changes reach production, and how operations are monitored, including what happens when the documented process is inconvenient.
Sample-based testing across the three domains, with evidence retained to external-auditor standards. Emergency changes and privileged accounts get specific attention because that is where control bypass concentrates.
Findings mapped to the financial impact they enable, distinguished by design versus operation, with a remediation sequence the IT function can actually execute.
This service is delivered on site and remotely across the firm's service area. See how it applies locally:
Not answered here? Ask Javed directly
They are the controls over the IT environment that everything else depends on, grouped into three areas: who can access systems and data, how changes to those systems are made and approved, and whether the systems run reliably day to day.
They matter because they are foundational. If anyone can change a report or a calculation without approval, no output from that system can be relied upon; which means an ITGC weakness contaminates every automated control and system-generated report above it, rather than affecting one account.
Yes, though the scope shifts. The vendor owns their change management and infrastructure (which is what their SOC report covers) but you own user access, tenant configuration, integrations, and the data moving in and out.
The item most often missed is the complementary user entity controls section of the vendor's SOC report. Those are controls the vendor states you must operate for their controls to be effective. Obtaining the report is not the control; satisfying the CUECs is.
Access reviews that are performed but not meaningfully performed. A manager receives a list of usernames and role codes they cannot interpret, approves it, and the control is documented as operating effectively.
The fix is not more frequent reviews. It is translating entitlements into plain descriptions of what a person can actually do, giving the reviewer someone to ask, and requiring an explicit response for each line rather than a single approval at the bottom.
They overlap heavily. SOC 2's common criteria include access control, change management and system operations, substantially the same ground as ITGC, evaluated against the Trust Services Criteria rather than against financial statement assertions.
Where a company needs both, testing once and reporting twice removes a large amount of duplicate work. Running them as separate engagements with separate documentation is common and wasteful. The SOC 2 page covers that side in detail.
Most of it, yes. Access listings, change records, configuration exports and evidence review are all handled remotely, and it is more efficient for everyone.
Walkthroughs are better in person for a first engagement, sitting with the person who actually grants access reveals things a screen share does not, particularly the informal workarounds that never appear in documentation. For clients in Thousand Oaks, Camarillo, Simi Valley and across Ventura County that is straightforward; for clients further afield it is usually a single visit at the start.
Scoping, control design, testing and remediation for Section 404, including the first-year programmes that decide whether year two is manageable.
SOX 404 supportNIST, ISO 27001, PCI DSS and CMMC assessment, with findings expressed as business exposure rather than a severity count.
Security assessmentArchitecture review, IAM and configuration assessment, and FedRAMP or FISMA readiness across the three major cloud platforms.
Cloud securityThree decades of audit, controls and finance leadership across banking, card, mortgage, insurance, staffing and semiconductor.
Get in touch
A sentence or two about your situation (the standard involved, the deadline, and what has already been attempted) is enough to get a useful reply.
Send the shape of it. The first call is diagnostic, not billed, and it regularly ends with a smaller engagement than the one you asked about.
Thirty years of audit, financial leadership and IT governance in one engagement, and a direct answer about scope, sequence and cost before anything is signed.
Or speak to Javed directly (310) 980-3958 Message on WhatsApp