Speaker: Business Engineering Corporation
Product Business Division
Nohara-san
Amidst growing calls for the importance of digital transformation (DX) in corporate management, business innovation in the finance and accounting fields is also accelerating. As many companies are shifting towards data-driven management, the efficient use of accounting data and advanced analysis have become essential.
GLASIAOUS, a cloud-based accounting and ERP system provided by Business Engineering Corporation, Ltd. (hereinafter referred to as B-EN-G), incorporates generative AI technology to improve the efficiency and sophistication of accounting operations. This time, we spoke with Mr. Nohara, the engineer leading the development, about the forefront of generative AI utilization and future prospects.
Implementing generative AI capabilities—the journey from prototype to production
Mr. Nohara, who is involved in the development of GLASIAOUS at B-EN-G, began working with generative AI about two years ago. He was in charge of exploring the possibilities of generative AI as part of a technical survey.
"I was skeptical at first," Nohara recalls. "Generative AI is sometimes talked about as a panacea, but I doubted whether it would actually be useful in a business setting. However, when I actually used it, I was surprised by its potential."
After the initial investigation, Mr. Nohara focused on its application to manual search functionality. His department has a system called the "10% Project," which allows developers to work on research and development on any theme they choose. He utilized this framework to begin developing a prototype manual search chatbot using generative AI.
"My experience creating the GLASIAOUS manual was extremely helpful during development. Because I had knowledge of what kinds of questions users might ask and which documents they should refer to, I was able to provide the AI with appropriate background information."
This prototype received high praise within the company and was officially adopted as a product feature, "GLASIAOUS Copilot," in just two to three months. Since then, further improvements have been made, and it is now used by many users.
Mr. Nohara points to "balancing product knowledge and technical knowledge" as the key to their success. "When implementing generative AI into business systems, simply calling APIs is not enough. It is important to understand the context of the business and provide the AI with appropriate information."
Development of functions that combine AI and OCR: Achieving multilingual support through generative AI
Following "GLASIAOUS Copilot," Mr. Nohara worked on developing a function that combines AI and OCR to read invoices. While OCR functionality had been a long-standing request from GLASIAOUS users, existing technologies presented challenges in terms of cost and multilingual support.
"GLASIAOUS is used not only in Japan but also overseas in countries like Thailand and Vietnam. For automated invoice processing, it's crucial not only to extract text from images but also to label information such as dates, amounts, and details. The most challenging aspect is dealing with country-specific rules. For example, Thailand uses its own Buddhist calendar, which differs from the Gregorian calendar by 543 years. We've been able to efficiently handle this complexity by leveraging the language processing capabilities of our generative AI."
Of course, a deep understanding of accounting operations and GLASIAOUS was key to developing this feature. "Without knowledge of what information on invoices is needed for accounting and how to link it with customer master data, the system would simply become a system that reads text," says Nohara.
Continuous data-driven improvements
GLASIAOUS's AI functionality is continuously being improved even after its release. Mr. Nohara particularly emphasizes a data-driven cycle of verification and improvement.
"We use data to check the accuracy and usage of AI-generated responses and identify areas for improvement. Developing AI functionality isn't a one-time release; continuous improvement is essential."
One of the initial challenges they faced was dealing with system-specific terminology. "While the generative AI is strong with general information on the web, it was weak with system-specific information such as GLASIAOUS's specialized terminology and function names. The solution we introduced is a method called 'hybrid search.' By combining generative AI with full-text search, we can now accurately answer questions that include specialized terminology." Following the release of this feature improvement, the number of users of the GLASIAOUS Copilot function has doubled.
Mr. Nohara says that his own development approach has also changed significantly thanks to generative AI. "Previously, I would get clues from searched information and then code, but now I can ask the generative AI questions and have it generate code that is almost ready to use. As a result, development speed has improved dramatically, and we can now implement more complex functions in a shorter period of time. I can proceed while confirming with the generative AI, 'Is this correct?', so my development approach has fundamentally changed. In development, it is also important to utilize AI and choose methods that are low in cost and highly effective."
However, Mr. Nohara also emphasizes that "AI alone will not suffice."
"Even if AI generates good code, it's humans who ultimately decide whether it fits the system, and whether there are any issues with performance or maintainability. The role of engineers may change, but their importance remains the same."
Accounting systems of the future: The challenge of "financial strategy support functions"
Currently, Ms. Nohara is working on developing a "financial strategy support function." This goes beyond simply automating accounting processes and aims to provide information that will be useful for management decisions.
"Current chatbots are strong at one-to-one question-and-answer interactions, but they are still not good at cross-system analysis or providing consulting-type advice. The financial strategy support function aims to have AI automatically detect financial warning signs such as worsening cost ratios or extended payback periods, and provide information to help decide whether to continue the project."
They say that achieving this requires deep learning that goes beyond simple question-and-answer sessions. "We are also considering developing an AI model that covers everything from GLASIAOUS's product knowledge and database structure to its API specifications."
Furthermore, in the future, we envision using AI in more specialized areas such as audit support and legal compliance. "Even if the final decision is made by a human, we believe that AI can significantly improve the quality and speed of decision-making by providing 'insights'."
Combining technology and business knowledge: the ideal engineer for the future
GLASIAOUS's AI function development case study suggests that the integration of technology and business knowledge will become increasingly important in system development in the era of generative AI. Mr. Nohara says the following about the qualities that engineers will need in the future.
"It's a misconception that anyone can develop products if they have AI. Rather, what's needed are people who can deeply understand the product's inner workings and design what kind of information to provide to the AI to obtain the appropriate results."
Furthermore, Mr. Nohara points out that "what to ask and how to ask it" is crucial when using AI. "Since AI cannot read between the lines like humans can, it is necessary to provide it with clear and concise assumptions. Whether or not you can design that is the key factor that determines whether an AI is usable."
Conclusion: Business transformation brought about by the synergy of AI and humans: The optimal fusion of technology and expertise
GLASIAOUS is driving the digital transformation of accounting operations by incorporating generative AI. Its distinguishing feature is that the introduction of technology itself is not the goal, but rather the primary focus is on solving the user's problems.
"AI is a tool, not an end in itself. It's important to address the fundamental challenge of streamlining and improving accounting operations and select the most suitable technology," Nohara emphasizes.
The evolution of generative AI technology is expected to accelerate in the future. While B-EN-G actively embraces cutting-edge technology, it maintains an approach that emphasizes a balance between human judgment and expertise. Integrating technology and business knowledge to contribute to solving our customers' fundamental challenges—that is the value of the DX that GLASIAOUS and B-EN-G provide.
