AI in Accounting and Auditing: Uses, Evidence, and Review

AI in accounting and auditing helps with the preparation and organization work that precedes professional judgment: reconciling large datasets, drafting work papers, summarizing disclosures, and identifying anomalies for an accountant or auditor to investigate. It does not apply professional judgment, assess the sufficiency of evidence, or issue an opinion. Tools like fenbs help accounting teams keep non-sensitive evidence references, review work, and workflow decisions organized across sessions without putting client or audit data in a shared AI tool.

The distinction between accounting preparation and audit evidence matters professionally. An accountant preparing financial statements applies professional standards and judgment to their client's data. An auditor assessing those statements must maintain independence and apply audit standards to evaluate whether the financial statements are materially correct. AI can assist with the analytical and organizational work in both functions, but the professional standards, the independence requirements, and the conclusions belong to the licensed professional.

AI in accounting preparation

Accounting preparation tasks are well-suited to AI assistance because they involve applying defined rules to structured data. Categorizing transactions, reconciling accounts, identifying exceptions, and drafting financial statement disclosures from templates are examples of work where AI can do the tedious part and present the output for review. A CPA then checks the output, applies judgment to edge cases, and approves the result.

Journal entry preparation, depreciation calculations, and deferred revenue schedules are also good candidates. The AI handles the calculation and the structured output; the accountant reviews for accuracy and appropriate accounting treatment. For anything that requires a judgment call, for example whether an expense should be capitalized or expensed, the AI can present the relevant standard and both treatments, but the accountant decides.

AI in audit support work

Audit support work includes document organization, preliminary analytics, and drafting work papers. AI can process a large volume of transactions and flag items that warrant attention: unusual amounts, transactions outside normal business hours, payments to unfamiliar vendors, or concentrations that differ from prior periods. Those flags become the starting point for audit procedures, not the conclusion.

Drafting work papers is another useful application. AI can produce a structured first draft of a work paper from notes and test results. The auditor then reviews for completeness, ensures the work paper supports the conclusion, and makes the professional determination about whether the evidence is sufficient and appropriate. The work paper is the auditor's document; the AI assists its preparation.

Independence is a foundational requirement of auditing. Auditors cannot use tools that compromise their independence or that give the appearance of compromising it. Before deploying any AI tool in an audit context, review it with your firm's independence policy and your professional standards. The AICPA's professional standards↗ (opens in a new tab) and the Public Company Accounting Oversight Board's standards apply to different engagements.

Evidence references and what belongs where

Audit evidence must be kept in the firm's authorized systems. Work papers, client documents, confirmations, and any information obtained during the engagement belong in the engagement file, under the firm's document retention policies and subject to applicable professional standards. That information should not be placed in a general AI tool.

What can be kept in a separate reference tool like fenbs: non-sensitive references such as accounting standards citations, audit standard references, firm methodology guidance, and approved procedure templates. The distinction is between the engagement information, which must stay in authorized systems, and the professional knowledge and procedures that support the work.

Professional review and what AI cannot replace

Professional judgment in accounting and auditing is irreplaceable by current AI. Materiality assessments, going concern evaluations, fair value measurements that require significant estimation, and the overall conclusion on whether financial statements are fairly presented require the professional's analysis, knowledge of the client and industry, and signature.

The history of the engagement also matters in ways AI does not automatically access. An auditor who has served a client for several years brings knowledge of that client's control environment, its management's integrity, and prior year findings that informs judgment on this year's work. AI working from a document does not have that context unless you provide it explicitly and appropriately.

How to do this with fenbs

1. Create a free account. Create a project for your accounting practice or audit team with the function and the standard procedures you want AI to assist with.
2. Add approved procedure templates as Skills and prompts: the work paper structures, disclosure templates, and analysis formats that your team uses across engagements.
3. Add verified standard references as Links: the applicable FASB ASC sections for accounting topics you handle frequently, the relevant PCAOB or AICPA standards for common audit areas, and the review dates.
4. Record review rules as Decisions and rules: which outputs require partner or manager review, which steps cannot be delegated, and which information must stay in the engagement file and not in fenbs.
5. Connect your AI assistant from Integrations and ask it to read the project context before starting any preparation or drafting session. Never input client data, engagement information, or audit evidence into fenbs.

fenbs does not store client data, financial statements, audit evidence, or personally identifiable information. It holds non-sensitive professional knowledge: standard references, approved templates, and review procedures. All engagement information stays in your firm's authorized engagement management system.