A new study shows that accountants are optimizing workflows with artificial intelligence, significantly speeding up closing processes and improving service quality. Research by Stanford Graduate School of Business Assistant Professor Jung Ho Choi and MIT PhD student Chloe Xie indicates that AI does not replace accountants but enhances overall efficiency by automating repetitive tasks, such as transaction classification, allowing accountants to focus on higher-value work.

Based on feedback from 277 accountants and task data from 79 small and medium-sized enterprises, the study found that accountants using generative AI can support more clients per week, complete monthly reports 7.5 days faster than traditional methods, and reduce back-office processing time by 8.5%. AI also helps accountants break down expenses into specific categories like bonuses and benefits, increasing report granularity by 12% and making financial information easier to analyze. "AI handles upfront tasks like information extraction and bank transaction linking, enabling accountants to serve clients more efficiently," Professor Choi explained.

However, AI benefits are closely tied to accountants' experience levels. Senior accountants tend to treat it as a partner, intervening when the system lacks confidence, while junior staff are more likely to accept uncertain outputs, resulting in limited performance gains. The study notes potential errors from AI-generated outputs in human-machine collaboration, requiring human review. Dr. Xie emphasized: "The technology is designed to augment expert capabilities, not replace humans. Complex judgment in accounting still requires human professional knowledge."

Despite 62% of accountants worrying about AI errors, 43% concerned about data security, and 37% fearing job stability, nearly half of respondents confirmed it improved task reliability and accuracy, with two-thirds citing automation of routine tasks as the greatest advantage. Currently, AI is mainly applied in basic areas like bookkeeping, with limited large-scale penetration in complex scenarios like auditing and taxation, though expansion is possible in the future. "AI can help auditors integrate information, but final judgments still require human decision-making," Professor Choi added.