Bill One Investigates "Working Styles of Accounting Departments" Focusing on AI Utilization
A survey conducted by Sansan Corporation targeting 2,112 accounting professionals revealed that over 60% do not use conversational AI in their work, primarily due to "anxiety about accuracy" and "security risks." AI utilization for accounting-specific tasks remains below 20%.
📋 Article Processing Timeline
- 📰 Published: May 2, 2026 at 00:00
- 🔍 Collected: May 1, 2026 at 15:32
- 🤖 AI Analyzed: May 1, 2026 at 15:52 (20 min after Collected)
Sansan Corporation, which provides AX services that change working styles, conducted a "Survey on the actual state of working styles in accounting" targeting 2,112 accounting department personnel.
The survey results showed that over 60% answered that they do not use conversational AI (*1) in their work. Even among those who responded that they utilize conversational AI, while it is used for highly versatile tasks regardless of job type, such as email creation and term search, its utilization in accounting-specific tasks such as financial data analysis/report creation, numerical reconciliation/consistency checks, and judgment based on internal regulations remained below 20% for all categories. The main barriers cited were "anxiety about accuracy" and "security risks," indicating the difficulty of AI utilization unique to accounting operations, which demand high accuracy and confidentiality.
■ Survey Result Summary
・Over 60% of accounting personnel do not use conversational AI.
・Even among those who stated they utilize conversational AI, while it is used for highly versatile tasks including email creation, its utilization in accounting-specific tasks remains below 20% for all categories.
・"Anxiety about accuracy" and "security risks" were cited as barriers to conversational AI utilization.
■ Background of the Survey
Recently, the utilization of AI in business has been spreading, but generative AI has been pointed out for issues such as hallucinations that output information not based on facts, and the risk of information leakage. Particularly in accounting operations, where information related to payments to business partners, tax payments, and financial closing is handled, high accuracy and confidentiality are required. Therefore, discerning which tasks to apply generative AI, which still carries risks, has become an important theme.
Thus, as Sansan Corporation, which provides the accounting AX service "Bill One," we conducted a survey on the current state of "working styles in accounting," including AI utilization.
■ Survey Results
・Over 60% of accounting personnel do not use conversational AI.
When 2,112 company employees in accounting departments nationwide were asked about their utilization of conversational AI in accounting operations, the most common response was "Interested but never used it" at 35.5%, followed by "Not interested and no plans to use it" at 20.3%. Furthermore, "Tried it in the past but not currently using it" was 4.4%, revealing that a total of 60.2% do not use conversational AI.
・Even among those who responded that they utilize conversational AI, while it is used for highly versatile tasks including email creation, its utilization in accounting-specific tasks remains below 20% for all categories.
When those who responded "Utilize conversational AI in their work" were asked about specific uses, "Creating/editing emails and chat messages" was the most frequent at 43.7%, followed by "General search/research for accounting/tax terms" at 39.0%, and "Summarizing/creating meeting minutes" at 33.6%.
On the other hand, "Financial data analysis/report creation" was 17.7%, "Reconciliation/checking of trial balances, financial statements, and bank statements" was 10.8%, and "Judgment of account titles/tax classifications and creation of journal entries" was 9.3%. While it is utilized for highly versatile tasks, its utilization in accounting-specific tasks such as financial data analysis/report creation, numerical reconciliation/consistency checks, and judgment based on internal regulations remained below 20% for all categories, revealing the current situation where the scope of applicable tasks is limited.
・"Anxiety about accuracy" and "security risks" were cited as barriers to conversational AI utilization.
When asked about the barriers to utilizing conversational AI in accounting operations, "Anxiety about accuracy" was the most common at 38.8%, followed by "Security risks" at 35.7%, and "Difficulty in standardization" at 33.0%. This result indicates the importance of accuracy and confidentiality due to the nature of accounting operations related to financial closing and payments, as well as the difficulty of standardizing work processes for AI utilization.
■ Comment from Aito Kasaba, VPoP (Vice President of Product), Bill One Division, Sansan Corporation
The survey results show that the utilization of generative AI remains limited. We believe the background to this is the nature of accounting operations, which are directly linked to financial closing and payments, where no mistakes are allowed. If generative AI were to be applied to actual operations, it would not be able to respond accurately according to each company's unique rules, and personnel would need to provide accurate and detailed instructions to the AI repeatedly. Furthermore, if even one mistake is found, personnel would have to re-verify all data output by the AI, which could even lead to a situation where "workload has not decreased" or "has even increased."
Against this background, Bill One, which advocates the tagline "create what can be eliminated, change the way all employees work."
The survey results showed that over 60% answered that they do not use conversational AI (*1) in their work. Even among those who responded that they utilize conversational AI, while it is used for highly versatile tasks regardless of job type, such as email creation and term search, its utilization in accounting-specific tasks such as financial data analysis/report creation, numerical reconciliation/consistency checks, and judgment based on internal regulations remained below 20% for all categories. The main barriers cited were "anxiety about accuracy" and "security risks," indicating the difficulty of AI utilization unique to accounting operations, which demand high accuracy and confidentiality.
■ Survey Result Summary
・Over 60% of accounting personnel do not use conversational AI.
・Even among those who stated they utilize conversational AI, while it is used for highly versatile tasks including email creation, its utilization in accounting-specific tasks remains below 20% for all categories.
・"Anxiety about accuracy" and "security risks" were cited as barriers to conversational AI utilization.
■ Background of the Survey
Recently, the utilization of AI in business has been spreading, but generative AI has been pointed out for issues such as hallucinations that output information not based on facts, and the risk of information leakage. Particularly in accounting operations, where information related to payments to business partners, tax payments, and financial closing is handled, high accuracy and confidentiality are required. Therefore, discerning which tasks to apply generative AI, which still carries risks, has become an important theme.
Thus, as Sansan Corporation, which provides the accounting AX service "Bill One," we conducted a survey on the current state of "working styles in accounting," including AI utilization.
■ Survey Results
・Over 60% of accounting personnel do not use conversational AI.
When 2,112 company employees in accounting departments nationwide were asked about their utilization of conversational AI in accounting operations, the most common response was "Interested but never used it" at 35.5%, followed by "Not interested and no plans to use it" at 20.3%. Furthermore, "Tried it in the past but not currently using it" was 4.4%, revealing that a total of 60.2% do not use conversational AI.
・Even among those who responded that they utilize conversational AI, while it is used for highly versatile tasks including email creation, its utilization in accounting-specific tasks remains below 20% for all categories.
When those who responded "Utilize conversational AI in their work" were asked about specific uses, "Creating/editing emails and chat messages" was the most frequent at 43.7%, followed by "General search/research for accounting/tax terms" at 39.0%, and "Summarizing/creating meeting minutes" at 33.6%.
On the other hand, "Financial data analysis/report creation" was 17.7%, "Reconciliation/checking of trial balances, financial statements, and bank statements" was 10.8%, and "Judgment of account titles/tax classifications and creation of journal entries" was 9.3%. While it is utilized for highly versatile tasks, its utilization in accounting-specific tasks such as financial data analysis/report creation, numerical reconciliation/consistency checks, and judgment based on internal regulations remained below 20% for all categories, revealing the current situation where the scope of applicable tasks is limited.
・"Anxiety about accuracy" and "security risks" were cited as barriers to conversational AI utilization.
When asked about the barriers to utilizing conversational AI in accounting operations, "Anxiety about accuracy" was the most common at 38.8%, followed by "Security risks" at 35.7%, and "Difficulty in standardization" at 33.0%. This result indicates the importance of accuracy and confidentiality due to the nature of accounting operations related to financial closing and payments, as well as the difficulty of standardizing work processes for AI utilization.
■ Comment from Aito Kasaba, VPoP (Vice President of Product), Bill One Division, Sansan Corporation
The survey results show that the utilization of generative AI remains limited. We believe the background to this is the nature of accounting operations, which are directly linked to financial closing and payments, where no mistakes are allowed. If generative AI were to be applied to actual operations, it would not be able to respond accurately according to each company's unique rules, and personnel would need to provide accurate and detailed instructions to the AI repeatedly. Furthermore, if even one mistake is found, personnel would have to re-verify all data output by the AI, which could even lead to a situation where "workload has not decreased" or "has even increased."
Against this background, Bill One, which advocates the tagline "create what can be eliminated, change the way all employees work."