AI-Powered SOX Automation for Computer Operations ITGC Testing
Computer Operations controls commonly form part of an organization’s IT General Controls framework when they support systems and processes relevant to financial reporting.
These controls help support the reliable processing, integrity, and, where relevant, availability of financial systems. They may cover activities such as batch processing, job scheduling, system monitoring, incident management, backups, and recovery procedures.
Weaknesses in these controls can result in incomplete processing, unresolved system issues, delayed financial reporting, or reduced reliance on automated controls and system-generated information.
The Traditional Testing Challenge
Testing Computer Operations controls is often highly manual and fragmented.
SOX teams may need to collect evidence from multiple sources, including job scheduling tools, system logs, monitoring platforms, backup reports, IT service management systems, and supporting documentation prepared by IT teams.
The evidence may be:
Distributed across multiple systems
Provided in inconsistent formats
Difficult to connect to the relevant control requirement
Reviewed manually through spreadsheets and screenshots
Limited to selected samples rather than broader populations
This process can require significant effort from SOX, IT, and business teams. It can also make it more difficult to identify recurring exceptions, confirm that the correct population was tested, and maintain a clear connection between the evidence, test procedures, and final conclusion.
AI-powered SOX automation can help organizations make this process more structured, consistent, and scalable.
Centralized Testing of ITGC
AI-powered SOX platforms can bring ITGC testing activities into a centralized testing environment.
Depending on the control design, available evidence, and supported data sources, testing may evaluate whether:
Financially relevant batch jobs completed in accordance with defined schedules
Operational incidents were identified, escalated, and resolved within required timeframes
Monitoring alerts were reviewed and addressed appropriately
Backup activities were completed according to defined requirements
Operational exceptions were investigated and resolved
Required reviews and approvals were documented
Automated Analysis of Operational Evidence
AI can support the collection, organization, and analysis of evidence used in Computer Operations testing.
Evidence may be collected through supported integrations, structured exports, and document uploads, depending on the systems used by the organization.
The platform can then help SOX teams:
Organize evidence from different systems and formats
Identify failed, delayed, or incomplete activities
Highlight unresolved incidents and recurring exceptions
Compare evidence against predefined testing criteria
Link identified exceptions to the relevant control and testing period
Prepare structured testing documentation for review
Depending on the control and the evidence available, testing may also consider factors such as job timing, reruns, dependencies, rejected records, processing exceptions, and evidence of issue resolution.
AI-assisted analysis can help interpret and organize unstructured evidence. Final testing results, however, should remain based on defined control criteria and traceable decision logic that reviewers can understand and validate.
Completeness and Accuracy of Information Used in Testing
Automated testing is only reliable when the information used in the test is appropriate for the testing objective.
SOX teams must consider whether the evidence represents the correct period, system, population, and data fields. They may also need to evaluate whether the information is complete and accurate before relying on it.
An AI-powered platform can support procedures designed to evaluate the reliability of information used in testing, based on the available source data and predefined validation criteria.
These procedures may include:
Reviewing report parameters and date ranges
Comparing record counts or totals to source information
Confirming that the relevant systems and activities are included
Identifying missing, duplicate, or unexpected records
Linking the evidence to the system or report from which it was generated
Documenting how completeness and accuracy were evaluated
The appropriate procedures will depend on the control, the source of the information, and the organization’s SOX methodology.
Moving Beyond Basic Status Checks
A successful status does not always demonstrate that an operational control functioned as intended.
For example, a batch job marked as completed may still require further evaluation to determine whether:
It ran during the required period
The correct data population was processed
Required dependencies completed successfully
Records were rejected or excluded
The job was rerun
Exceptions were identified and resolved
The issue had a potential impact on financial reporting
AI-powered automation can help apply these considerations consistently where the required information is available.
This allows SOX teams to move beyond basic status reviews and perform testing that is more closely aligned with the control objective.
Broader Population Analysis
Traditional testing often relies on a sample of operational activities.
Where a complete and reliable population is available, automation can apply defined testing procedures across a broader population, and in some cases the full population, of relevant jobs, incidents, backups, or other operational events.
Broader population analysis can help:
Increase testing coverage
Identify recurring issues
Detect exceptions that may not appear in a limited sample
Apply consistent criteria across the testing period
Provide greater visibility into the frequency and nature of control failures
Full-population analysis does not eliminate the need for professional judgment. SOX teams must still evaluate the control design, the reliability of the underlying information, the significance of exceptions, and whether all relevant elements of the control can be tested using the available data.
Explainable Testing and Traceability
SOX testing requires more than an automated result.
Reviewers need to understand:
What was tested
Which evidence was used
Which criteria were applied
What exceptions were identified
How those exceptions were evaluated
How the final conclusion was reached
walktru combines AI-assisted evidence analysis with predefined testing criteria and traceable decision logic.
Through SOX TRAIL®, relevant evidence, testing procedures, identified exceptions, review activity, and conclusions can be connected within a traceable testing record.
This allows reviewers to follow the testing process from the control requirement to the supporting evidence and final conclusion. It also helps organizations maintain structured, reviewable, and audit-ready documentation.
Value Across the Organization
AI-powered automation can create value for the different stakeholders involved in ITGC testing.
SOX teams
Reduced manual evidence handling, more consistent testing procedures, broader coverage where appropriate, and clearer documentation of testing results.
IT operations teams
More structured evidence requests, improved visibility into recurring operational issues, and less time spent responding to repetitive follow-up questions.
Finance management
Greater visibility into how operational controls supporting financial reporting are tested and how identified exceptions are evaluated.
External auditors
More structured and traceable testing documentation that can support efficient review when the evidence and testing approach meet the auditor’s requirements.
Conclusion
TGC testing plays an important role in supporting reliable financial reporting, but the testing process has traditionally been manual, fragmented, and difficult to scale.
AI-powered SOX automation can help organizations centralize evidence, apply predefined testing procedures, analyze broader populations, identify exceptions, and maintain a traceable connection between control requirements, evidence, testing activity, and conclusions.
walktru helps SOX teams move beyond managing the testing process to executing more of the work itself, while preserving the accuracy, traceability, explainability, and accountability required in highly controlled environments.