Artificial intelligence (AI) is one of the most discussed technologies today, with seemingly limitless capabilities. Accounting firms are actively adopting AI tools to support their day-to-day operations, particularly within audit engagements.
Artificial intelligence in auditing can help teams process information more efficiently and identify areas that may need additional attention. It also introduces risks and limitations that auditors need to understand and manage, while keeping professional judgment at the center of the engagement.
The benefits of AI in an audit
The introduction of AI into the audit process has transformed how auditors can spend their time. One of its most significant benefits is its ability to process large volumes of data and identify information that may be meaningful based on prompts provided by the auditor.
Analyzing large volumes of data
For example, an auditor may need to identify unusual or potentially high-risk journal entries within a client’s general ledger. Traditionally, this could require reviewing thousands of transactions and manually extracting entries that warrant further investigation.
AI can assist in analyzing this data and, within minutes, produce a smaller population of transactions for the audit team to evaluate. Tasks that previously required hours, or even an entire day, can often be completed in a matter of minutes.
By processing data more quickly, AI allows auditors to redirect time previously spent on manual tasks toward higher-value activities. In artificial intelligence in auditing, that benefit depends on the quality and completeness of the information being analyzed, as well as the auditor’s review of the results.
The tool’s output is a starting point for analysis. The auditor still needs to understand what information was included, how the tool was used, and whether the results are relevant to the audit objective.
Making more time for professional judgment
While the resulting time savings are valuable, the greater opportunity lies in how that time can be used. Rather than spending most of the day identifying transactions, auditors can devote additional time to discussing those transactions with management and gaining a deeper understanding of why they occurred.
They can ask questions, evaluate the underlying processes, and determine whether the transactions indicate a broader issue within the client’s accounting practices or internal control environment. The additional time can support a more informed conversation about the circumstances behind an unusual item.
This is where AI can help auditors move beyond simply identifying observations and toward providing meaningful, value-added recommendations. Artificial intelligence in auditing may help auditors find patterns or exceptions, while the people on the engagement team assess what those results mean for the client.
AI may also help organize information or compare records so auditors can focus their attention on questions that require judgment. These uses can support the audit process, but they do not determine whether a transaction is appropriate or whether a control is operating effectively.
The potential downsides of AI in an audit
Despite its potential benefits, AI also has significant limitations. Perhaps the most obvious is its ability to generate incorrect responses or produce results that sound convincing but do not adequately address the facts.
Incorrect or incomplete responses
AI tools can process auditing standards and other technical information, making them useful resources for auditors. For example, an auditor may ask an AI tool about the appropriate accounting treatment for a particular transaction and receive a response that appears confident and well supported.
However, the answer may still be incorrect or may fail to consider an important fact, exception, or nuance. A response can be clearly written without being complete, accurate, or appropriate for the specific engagement.
This creates a significant risk for auditors. If an auditor relies on an incomplete answer without checking the underlying information, the audit team may overlook a relevant issue or spend time following a result that does not fit the client’s circumstances.
Artificial intelligence in auditing should therefore be treated as a way to assist analysis, not as an authority on accounting or auditing requirements. Auditors need to assess whether a tool’s response is supported by reliable information and whether it addresses the question that was actually asked.
Verification and professional skepticism
It is imperative that auditors critically evaluate and verify any information generated by AI. AI should be viewed as a tool that supports the auditor’s work, not as a substitute for professional judgment.
Auditors need to evaluate information that supports a conclusion as well as information that contradicts it. They also need to consider whether the evidence is sufficient, relevant, and reliable for the purpose of the audit.
When an AI tool identifies unusual journal entries, for example, the audit team should investigate the items and understand the circumstances behind them. A flagged transaction is not, by itself, proof of an error or fraud. A transaction that was not flagged is not, by itself, proof that no issue exists.
Ultimately, responsibility for audit conclusions remains with the auditor, regardless of how information was obtained. The audit preparation guidance from Bowers also underscores the value of organized documentation, effective controls, and communication during an audit.
Confidentiality and data handling
Auditors also need to understand how information entered into an AI tool is handled. Before using a tool with client information, the engagement team should follow the firm’s policies and understand the tool’s access, storage, and data-use settings.
Client records may contain confidential financial, operational, or personal information. Artificial intelligence in auditing should be used with appropriate safeguards so the efficiency of a tool does not come at the expense of confidentiality or the firm’s responsibilities to the client.
Teams should use approved tools for the intended purposes and be thoughtful about what information they enter. If a tool cannot be used with sensitive client information under firm policy, auditors should not enter that information into it.
AI can affect both the auditor and the client
Artificial intelligence in auditing involves more than an audit firm’s use of AI tools. A client may also use AI within accounting, financial reporting, or controls, and that use may affect the information the auditor examines.
When the audit team uses AI
When an audit team uses AI, the auditors should understand the task the tool is performing and how the output will be used. They need to evaluate whether the tool is suitable for the procedure and whether the information supplied to it is complete and accurate.
The team should also consider how the work will be reviewed and documented. An experienced auditor reviewing the work should be able to understand the procedure performed, the results produced, and how the results support the audit conclusion.
That may require retaining information about the data set, the tool’s role, and the auditor’s review. If an output is used as audit evidence, the team should be able to explain why it is relevant and reliable for that purpose.
AI can help identify where an auditor may want to look more closely. The auditor must still determine whether additional procedures are needed and whether the evidence gathered supports the conclusion.
When the client uses AI
A client’s use of AI in financial reporting can raise different questions. Auditors may need to understand where the client uses AI, what information the tool processes, and whether the output affects accounting records, financial statements, or internal controls.
That understanding can help the auditor assess whether the client’s processes introduce risks that should be considered in planning the audit. The auditor may also need to understand how the client reviews AI-generated information and responds when an output appears incorrect.
For example, if a client uses AI to process accounting information, auditors may need to understand the data involved and the controls over how the information is generated and reviewed. The auditor’s work remains focused on obtaining appropriate evidence and forming an independent conclusion.
For additional context about AI and audit readiness in financial reporting, see EY’s discussion of AI audit readiness. The related questions about controls, evidence, and review apply whether AI is used by the organization being audited or by the audit team.
Using AI to support audit quality and client value
AI can help teams handle repetitive, manual, and data-intensive tasks, but technology alone does not create a high-quality audit. The value comes from using the time and information it provides to support careful analysis, communication, and follow-up.
Turning observations into useful questions
When AI identifies an unusual item, the auditor can use that result to guide follow-up questions. The team can discuss the transaction with management, understand the business reason for it, and consider whether the explanation is consistent with other information obtained during the audit.
Those conversations may help auditors understand whether an observation is isolated or connected to a broader process. They may also help the team determine whether a control, accounting practice, or documentation process should receive additional attention.
This is where artificial intelligence in auditing may support more meaningful recommendations. AI can help surface information for review, while auditors use their knowledge of the client and the engagement to decide what action, if any, is appropriate.
The objective is not to produce more observations for their own sake. It is to help the audit team understand the information it has gathered and communicate relevant findings clearly.
Building oversight into AI use
Audit firms need clear practices for selecting and using AI tools. Engagement teams should understand which tools are approved, what tasks they are intended to perform, and who is responsible for reviewing the results.
Training is also important. Auditors need enough familiarity with the tools they use to recognize their limitations, ask useful questions, and identify when a result needs further verification or escalation.
Professional judgment and supervision remain central. A useful way to think about artificial intelligence in auditing is that the tool may assist with a task, but the engagement team remains responsible for how that task is planned, performed, reviewed, and documented.
For a broader perspective on how AI may shape internal audit work, read Deloitte’s insights on GenAI’s risks and rewards for internal audit. The article discusses the opportunities AI can create alongside the need to evaluate its risks.
Keeping communication part of the process
Technology can change how information is analyzed, but communication between the auditor and management remains important. Auditors need to ask questions, understand the explanations they receive, and follow up when information is unclear or inconsistent.
Clear communication can also help the client understand why the audit team is asking about a process or transaction. The auditor can explain what information is needed and how it relates to the work being performed.
Bowers’ Audit & Assurance services include independent audits, reviews, compilations, and internal control reviews. The engagement’s scope and applicable standards determine which procedures are appropriate.
The future of auditing
AI is likely to continue changing the way audits are performed. Its most significant impact may not be the elimination of audit procedures, but rather the redistribution of an auditor’s time.
More time for judgment and understanding
If AI can help auditors spend less time on repetitive, manual, and data-intensive tasks, they can devote more time to exercising professional judgment, communicating with management, understanding business processes, and identifying opportunities for improvement.
Artificial intelligence in auditing may allow teams to examine information differently and focus follow-up on areas that warrant attention. Those possibilities still depend on sound data, thoughtful use of the tool, and a careful assessment of the results.
The future of auditing may therefore be less about AI replacing auditors and more about AI enabling them to become more valuable advisors to their clients. The auditor remains responsible for the work, the professional judgment applied, and the conclusions reached.
Keeping people accountable for the conclusions
As AI tools change, audit teams will continue to evaluate where the tools are useful and where human review is essential. A tool may help organize or analyze information, but the auditor must decide what the information means and what additional work is needed.
That responsibility is central to the profession. Artificial intelligence in auditing can support efficiency and help teams make better use of their time, while the auditor remains accountable for the quality and integrity of the audit.
FAQ
These questions summarize the opportunities and limitations of using AI in audit engagements.
How can AI help auditors?
AI can help auditors process large volumes of information and identify transactions or other items that may warrant additional attention. This can free up time for analysis, discussions with management, and professional judgment.
Can auditors rely on AI-generated answers?
No. Auditors need to critically evaluate and verify AI-generated information. A response may sound confident while still being incorrect or missing important facts or context.
Does AI replace an auditor’s professional judgment?
No. AI can support the auditor’s work, but responsibility for evaluating evidence and reaching audit conclusions remains with the auditor.
Can a client’s use of AI affect an audit?
Yes. A client’s use of AI in accounting, financial reporting, or internal controls may affect the information the auditor examines and the auditor’s understanding of relevant processes.
What is the main risk of using AI in an audit?
One significant risk is relying on an incorrect or incomplete output without verification. Auditors also need to consider the information entered into a tool and follow the firm’s policies for handling client data.