Generative Artificial Intelligence and Its Impact on the Utility Accounting Department

What Generative AI Actually Does

Generative AI — the technology behind ChatGPT, Claude, Microsoft Copilot, and similar tools — produces new content in response to prompts. Unlike traditional software that follows explicit rules, these tools generate text, analysis, and explanations by drawing on patterns learned from vast training datasets. For utility accounting professionals, this creates a genuinely new kind of productivity tool: one that can draft, explain, summarize, and analyze in natural language.

High-Value Applications in Utility Accounting

Document drafting is the most immediate application: AI can produce first drafts of board financial narratives, variance explanations, regulatory filing sections, audit response letters, and accounting policy memos in a fraction of the time it takes to write from scratch. The accountant's role shifts from writing to reviewing and refining — a more efficient use of expert knowledge.

Regulatory and standards research is another strong use case. Generative AI can explain GASB pronouncements, summarize FERC accounting guidance, or walk through the application of ASC 980 to a specific transaction. This is valuable for training and for quickly orienting staff to new standards, though the output should always be verified against actual source documents.

The Hallucination Problem: Why Human Review Is Non-Negotiable

Generative AI has a well-documented limitation: it sometimes produces confident-sounding information that is simply wrong. An AI might cite a FERC account number incorrectly, misstate the criteria for a regulatory asset, or fabricate a GASB provision. These errors look plausible and are easy to miss without careful review.

The solution is not to avoid AI — it is to treat AI output as a first draft that requires expert review before use. Establish a clear practice: AI drafts, accountants verify. Never submit an AI-generated regulatory filing, board report, or journal entry without human review against the underlying standards and data.

Data Security Considerations

Public generative AI platforms may use conversation data for model training. Before entering any utility financial data, customer information, or confidential materials into these platforms, verify your organization's AI security policy and consider enterprise licensing that provides data privacy protections. The productivity gains from AI are real, but not at the cost of inadvertent data disclosure.