Julia Chirkova
The integration of Artificial Intelligence (AI) into the auditing profession presents a transformative opportunity, not to replace human expertise, but to significantly enhance efficiency and focus by automating tedious administrative tasks. This perspective, shared by Shannon Smith, founder and CEO of Yak Tech and a former SOC auditor, challenges the prevailing notion that AI’s primary role is to supplant auditors. Instead, Smith advocates for a strategic deployment of AI to handle repetitive, structured, and administrative workloads, thereby freeing up auditors to concentrate on higher-value activities that demand professional judgment, critical thinking, and accountability.
The historical landscape of auditing is often characterized by an overwhelming volume of manual processes. For auditors like Smith, the daily grind involved a relentless stream of administrative duties: dispatching countless client emails, meticulously reconciling spreadsheets, manually copying and updating documents, and tracking requests across disparate, non-integrated tools. This labor-intensive environment not only consumed valuable time but also created fertile ground for errors and inefficiencies. The cyclical nature of audits meant that tasks were often repeated year after year, with controls copied from prior reports, updated, sent to clients, and then the returned information painstakingly reconciled against the changes. These ingrained inefficiencies could even lead to lingering stress, with Smith recalling occasional "nightmares" about the administrative burdens during peak audit seasons, even years after leaving the auditing profession.
Beyond the core audit work, the cleanup phase was equally arduous. Spreadsheets would crash, necessitating the re-entry of data. A single change in one document could trigger a cascade of manual updates across multiple other files, underscoring a profound lack of interconnectedness and automation in existing workflows. This backdrop of operational friction and manual redundancy is the critical context for understanding the current debate surrounding AI in auditing.
The Strategic Role of AI in Auditing: Automating the Administrative Burden
Smith’s core argument centers on a fundamental question: "What work should technology take off an auditor’s plate, and what responsibility needs to stay with the auditor?" This inquiry pivots the conversation from a superficial adoption of AI for its own sake to a thoughtful integration designed to address specific pain points.
A(I) is for Admin: Streamlining Repetitive Tasks
The most immediate and impactful application of AI and automation within the audit process lies in tackling repetitive, structured, and administrative tasks. These are the tasks that, while essential, consume significant auditor time, increase the potential for human error, and divert attention from the analytical and judgmental aspects of an audit.
AI’s capabilities can be leveraged to:
- Organize Evidence: Automatically categorizing and filing audit evidence, reducing the manual effort required to manage vast amounts of documentation.
- Compare Documents: Efficiently identifying discrepancies and similarities between different versions of documents, such as client contracts, financial statements, or policy documents.
- Identify Changes: Flagging modifications in documents or data sets, which is crucial for tracking updates and ensuring consistency across audit workpapers.
- Flag Missing Information: Proactively identifying gaps in required documentation or data, allowing auditors to request necessary items promptly.
- Prepare Follow-Up Requests: Generating standardized follow-up requests to clients based on identified information gaps or discrepancies.
- Generate Documentation: Automating the creation of routine audit documentation from existing information within the audit workflow, thereby reducing the need for manual report generation.
A practical illustration of this lies in the management of controls that are mapped to multiple regulatory requirements or appear across various reports. Traditionally, an auditor would test such a control once and then ensure that the relevant information was consistently reflected in all applicable documentation. If the original control was modified, the auditor would then face the tedious task of updating every instance across all related documents. Technology, particularly AI-powered automation, is ideally suited to manage this structured workflow, ensuring that changes are propagated accurately and efficiently across all relevant touchpoints.
Similarly, tasks such as reconciliations and initial reviews can be significantly augmented by AI. AI systems can swiftly identify discrepancies, compare data sets, and surface items that warrant further auditor attention. In essence, AI can serve as an invaluable "second set of eyes," provided that the auditor has full visibility into its review process and can verify its findings.
The Underpinning of Efficiency: Data Management and Workflow Automation
The efficiency gains promised by AI in auditing are intrinsically linked to improved data management and workflow automation. Historically, the lack of interoperability between audit tools created significant bottlenecks. Information siloed in different systems necessitated manual data transfer and reconciliation, a process prone to errors and delays.
A report by the Association of International Certified Professional Accountants (AICPA) in 2022 highlighted that auditors spend an average of 30% of their time on administrative tasks. This statistic underscores the scale of the inefficiency that AI aims to address. By automating tasks like data extraction, document comparison, and report generation, AI can reclaim this significant portion of an auditor’s time.
For instance, consider the process of gathering client data. Traditionally, this involved numerous emails, follow-ups, and manual data entry into various spreadsheets and audit software. With AI-powered tools, clients could upload documents to a secure portal, and AI could then automatically extract relevant data, categorize it, and populate the audit workpapers. This drastically reduces the manual effort for both the auditor and the client, leading to faster audit cycles and improved client satisfaction.
Furthermore, the "chain of evidence" in auditing is paramount. AI can enhance this chain by ensuring that all data inputs and transformations are logged and traceable. When AI performs a reconciliation, for example, it can generate an audit trail detailing the source data, the logic applied, and the resulting discrepancies. This transparency is crucial for auditor review and for defending the audit findings.
The Indispensable Human Element: Judgment, Accountability, and Professional Skepticism
While the potential for AI to streamline administrative tasks is undeniable, Smith strongly cautions against overestimating its role in making final audit conclusions. The crucial distinction, she emphasizes, is between performing a task and owning the outcome.
H(uman) is for Hold Up: The Limits of AI in Audit Conclusion
Many technology providers entering the audit space, Smith observes, tend to blur this line. While an AI system might adeptly analyze a document, flag a potential exception, or even suggest that certain evidence meets a requirement, it should not independently conclude that a control has passed, deem evidence sufficient, dismiss an exception, or issue a final audit conclusion without robust human oversight.
Auditing is inherently a process that relies on a delicate balance of evidence, professional judgment, and accountability. The auditor must possess a comprehensive understanding of:
- What was tested: The scope and nature of the audit procedures performed.
- Which evidence was used: The specific data and documentation relied upon for conclusions.
- Period and population coverage: Whether the evidence adequately represents the relevant time frame and the entire population under review.
- Assumptions made: The underlying assumptions that guided the audit process.
- Exception handling: How any identified exceptions or anomalies were investigated and resolved.
A mere "final approval button" on an AI-generated report does not equate to a human-led process. The auditor must retain the prerogative to critically evaluate the AI’s output, challenge its findings when necessary, and ultimately bear responsibility for the audit conclusion.
This is particularly vital because audit work is not always as standardized as technology developers might assume. Different audit firms develop unique methodologies and approaches to assessing whether evidence adequately addresses risks and control objectives. The concept of "Bring Your Own Methodology" (BYOM), which Smith advocates for, highlights the need for technology to support these diverse firm-specific approaches rather than imposing a singular, prescriptive process. AI tools should be flexible enough to adapt to these varied methodologies, rather than forcing firms to conform to a one-size-fits-all technological solution.
The Audit Trail is Not Optional: Ensuring Transparency and Trust
The foundational principle of auditing is the ability to defend the work performed. Handing over the entire process to an AI agent would undermine this critical tenet. When an AI tool generates a recommendation or a finding, the auditor must be able to trace the origin of that conclusion. This involves understanding:
- The information that informed the AI: The specific data points and documents the AI analyzed.
- The underlying evidence: Access to the raw data and audit evidence.
- How the result was reached: The logical steps and algorithms employed by the AI.
A system that merely accepts input and produces a polished report, while visually impressive in a demonstration, is insufficient for the rigorous demands of an audit. Auditors require deep visibility into the evidence, the process, and any exceptions encountered.
Trust in the audit process hinges on integrity and traceability. When information enters an AI system, the auditor must have confidence that it remains accurate and can be traced throughout the entire workflow. Any unexplained alteration to data, controls, or testing notes should trigger immediate scrutiny and investigation. This ensures that the audit trail remains robust and that the integrity of the audit process is maintained.
The Future of Audit: A Human-Centric Approach Enhanced by AI
The ultimate goal, Smith asserts, is not to create a less human audit process, but a more effective and human-centric one. AI and automation are not replacements for professional skepticism, which remains an indispensable attribute of every auditor.
Use the Tool, but Keep the Humanity: Augmenting Auditor Capabilities
Smith’s vision, which underpins the development of AI capabilities within Yak Tech’s infrastructure, is to empower auditors by freeing them from the drudgery of administrative tasks. This allows them to dedicate more time and cognitive energy to:
- Investigating Risks: Deeper dives into identified risks and potential vulnerabilities.
- Understanding Exceptions: Thoroughly analyzing and interpreting anomalies.
- Communicating with Clients: Engaging in more meaningful dialogue and collaboration with clients.
- Applying Professional Judgment: Exercising critical thinking and making informed decisions based on complex factors.
Furthermore, leveraging AI infrastructure can address the persistent capacity challenges faced by audit firms. Simply hiring more staff to manage inefficient workflows can lead to increased costs without a proportional increase in effective capacity. Developing an auditor capable of independently recognizing risk, evaluating unusual situations, and making sound judgments requires years of experience and dedicated training. AI can accelerate the development of these skills by allowing junior auditors to focus on analytical tasks earlier in their careers, supported by AI-driven insights.
However, Smith firmly rejects the notion that AI should remove the human element from the audit process. Unlike some companies that advocate for replacing auditors with AI, Smith believes that auditors serve a critical role beyond simply identifying client errors. Their function encompasses objective evaluation of evidence, clear communication of findings, and guiding organizations toward process improvements.
The most effective application of AI in auditing, therefore, is not to render the auditor obsolete. Instead, it is to meticulously remove the mechanical, repetitive work that hinders auditors from performing their jobs with optimal effectiveness. By strategically integrating AI to handle administrative burdens, the auditing profession can unlock new levels of efficiency, accuracy, and strategic insight, ultimately strengthening the integrity and value of audit services. This approach ensures that technology serves as a powerful enabler, augmenting human expertise and fostering a more robust and trust-worthy audit landscape.
About the Author:
Shannon Smith is the founder and CEO of Yak Tech, an automation and artificial intelligence platform for audit and accounting firms, and a former SOC auditor.
Photo Credit: Freepik
