Generative AI Research and Service Commercialization
First written in FY2023. 7 of 9 entries say "Advanced", first in FY2024-H1. Latest entry: FY2026-H1.
4 other companies' latest entry also says "Restated".
- FY2026-H1RestatedAnnual report
「It is essential to develop services that are practical for accounting operations through the use of generative AI. We recognize that a key future challenge for the Group is to continue generative AI research and development and connect its results to the provision of higher-value-added new services.」
The company maintains its policy of continuing generative AI research and development and linking the results to higher-value-added new services. R&D expenses for the interim consolidated accounting period were ¥99,714 thousand. - FY2025AdvancedAnnual report
「We are currently conducting basic research on generative AI and aim to establish technologies that will form the components of advanced services based on the results of this research in the future. We have already released some services using generative AI.」
The company is developing an LLM and AI agent with strong capabilities in bookkeeping and financial accounting and has released some generative AI-based services, with implementation at existing customers. R&D expenses were ¥198,340 thousand. Research resources are focused on improving accounting-specific LLMs and AI agents and implementing them in products rather than submitting papers to international conferences. - 2026-02-27AdvancedManagement Planning
「Commercialized as an accounting AI agent and conducted POCs with multiple customers, confirming a major contribution to business automation. Also applied to the new lease accounting standard product.」
The company has progressed from describing generative AI and LLM research to commercializing an accounting AI agent, conducting POCs with multiple customers and applying it to its new lease accounting standard product. - FY2025-H1AdvancedAnnual report
「During the interim accounting period, our R&D activities focused on developing an LLM with strong capabilities in bookkeeping and financial accounting for accounting operations. We also worked on additional development to improve the reading accuracy of existing receipts and invoices, expanded ancillary services such as line-item reading, and developed algorithms to expand the range of documents that can be read, including foreign-language supporting documents. R&D expenses for the interim accounting period were ¥59,141 thousand.」
FA Research continues its generative AI and LLM research, including development of an LLM specialized in bookkeeping and financial accounting, improvements in reading accuracy, line-item reading, and expanded reading of supporting documents, including foreign-language documents. R&D expenses were ¥59,141 thousand. - FY2024AdvancedAnnual report
「We have also begun adding functions to accounting systems provided by accounting software vendors and launched a service that supports decision-making in accounting operations using generative AI.」
Building on the FA Research structure, the company is developing accounting-focused LLMs and multimodal LLMs, improving reading accuracy, supporting foreign-language documents, and has launched a generative AI-based accounting decision-support service. R&D expenses were ¥94,658 thousand. - FY2024-H1AdvancedAnnual report
「To improve AI-OCR technology specializing in accounting documents such as receipts and invoices and make routine accounting operations more efficient, we are conducting research and development on the latest technologies, including generative AI and LLMs (Large Language Models). We have established an internal research team, FA Research, comprising AI researchers with doctorates in computer science and related fields.」
The company established the FA Research team and has advanced concrete research and development involving generative AI, LLMs, and multimodal LLMs. R&D expenses for the interim accounting period were ¥32,713 thousand, covering improvements in reading accuracy, support for foreign-language documents, and document analysis. - FY2023First statedAnnual report
「We will conduct research and development related to generative AI using LLMs (Large Language Models) and other technologies and work to develop high-value-added services.」
The company plans to conduct basic research on generative AI using LLMs, improve OCR accuracy, automate sophisticated reconciliation, and eventually commercialize services for operational automation and decision support.