Strengthening the Sales and Deployment Support System for the Enterprise Generative AI Platform "AskDona®"

To Whom It May Concern
  • GFLOPS Co., Ltd.
  • RIKEN Suuri Co., Ltd.
  • JSOL Corporation

Strengthening the Sales and Deployment Support System for the Enterprise Generative AI Platform "AskDona®"

Advancing generative AI adoption and social implementation across companies, research institutions, and public bodies such as local governments

GFLOPS Co., Ltd. (“GFLOPS”), RIKEN Suuri Co., Ltd. (“RIKEN Suuri”), and JSOL Corporation (“JSOL”), through the enterprise generative AI platform “AskDona®,” will support the advancement of business processes at companies, research institutions, and public bodies such as local governments. In areas such as leveraging business knowledge, handling inquiries, and assessment work, AI organizes answer drafts and first-pass evaluations grounded in evidence, freeing staff to concentrate on the work that only people can do, such as filling in the gaps and weighing countermeasures.

By combining their technical foundations, expertise in applying AI to specialized domains, and know-how in business deployment and operational support, the three companies will provide proposals and deployment support tailored to each customer’s challenges and stage of adoption, aiming to expand the practical application and social implementation of generative AI. Through this initiative, customers can advance generative AI adoption with accuracy, evidential grounding, and business fit, receiving the support they need at each stage, from technical validation to deployment design, business application, and embedding it into day-to-day operations.

Background

In recent years, the business use of generative AI has expanded beyond mere text generation and search support to a stage where it leverages the documents, regulations, and knowledge accumulated within an organization as evidence to support business decisions and specialized work. At the same time, for companies, research institutions, and public bodies such as local governments to adopt generative AI in practice, they must address multiple challenges: accuracy, evidential grounding, security, integration with existing operations, and making it stick in day-to-day operations after deployment.

To meet these market needs, GFLOPS, RIKEN Suuri, and JSOL have established a sales agency structure centered on “AskDona®,” the enterprise generative AI platform developed and provided by GFLOPS. Rather than establishing a specific joint venture (JV), this initiative builds a parallel, mutually complementary cooperative relationship in which each company specializes in its own expertise and strengths. This accelerates progress toward a society where “AI handles what can be left to AI, and people handle what only people can do.”

AskDona® product features and results to date

AskDona® is an enterprise generative AI knowledge platform for business domains where accuracy and evidential grounding are especially required. Combining RAG (Retrieval-Augmented Generation), AI agents, and large language models, it supports answer generation, inquiry handling, research, and assessment work, grounded in the documents and knowledge held by companies and research institutions.

(1) AskDona® core (the base platform)

The AskDona® base platform ingests documents and data across a wide range of formats, then searches and organizes that information to generate answers grounded in the sources it references. It doesn’t require you to prepare data document by document or tune it individually up front; instead, its proprietary architecture lets you put large volumes of documents to work as knowledge.

Results (RIKEN, supercomputer Fugaku): AskDona® was introduced on the support site of the supercomputer Fugaku in July 2024, and in February 2025 the transition of first-line inquiry handling was fully completed. It handles an average of about 680 inquiries per month and recorded a high accuracy of 83 points on a complex-question benchmark, successfully reducing the workload for handling inquiries by up to 61%.

(2) AskDona® “Batch Assessment” feature

Batch Assessment is a feature in which, for a predefined list of questions or evaluation items, an AI agent reads across a specified set of documents and, for each item, searches for the relevant information, evaluates it, organizes the reasoning behind the judgment, and presents evidence and referenced sources, all in bulk, to support first-pass evaluation.

It can be used for work with many check items where the necessary information is scattered across multiple documents, such as risk assessment, audits, compliance checks, and responding to security checklists.

Results (joint validation with JSOL): From April to October 2025, a joint validation of system risk assessment automation was conducted across all of JSOL. Registering the SMBC group’s 400-plus risk-evaluation items and design/operation documents for more than 100 systems into a dedicated RAG, the validation achieved over 90% accuracy in the four-way classification. As a result, JSOL achieved a company-wide reduction of roughly 2,000 hours (about 270 person-days) per year, halving the time from 4,300 hours to 2,300 hours.

Reference: December 22, 2025 news release, “Providing a solution to streamline system risk assessment using generative AI” https://www.jsol.co.jp/release/2025/251222.html

Technological innovation in generative AI is advancing at a remarkable pace, and its use is no longer optional but is becoming an essential management priority for every organization to tackle. At the same time, because generative AI does not possess an organization’s specific information, a mechanism that bridges an organization’s knowledge and generative AI is indispensable for it to deliver value in practice. AskDona® is precisely the platform that plays this role, and it delivers reliable results especially in business domains that demand accuracy and evidential grounding and where failure is not an option.

By bringing together three companies with complementary strengths, we aim to break the impasse in which generative AI adoption stalls short of expectations, and to support the transformation of as many organizations as possible. Adopting AskDona® is not merely about operational efficiency; it is an investment in an organization’s future that concentrates talent on “the work only people can do.” Companies, research institutions, and public bodies such as local governments that are serious about the practical use of generative AI are encouraged to contact each company’s point of contact.

About the companies

GFLOPS Co., Ltd.

Head office: Shibuya-ku, Tokyo / CEO: Maria Morimoto / https://gflops-ai.com

GFLOPS provides enterprise generative AI solutions under the philosophy of “let AI handle the work that can be left to AI, so people can focus on the work only they can do.” Aiming to build a secure foundation that can handle even highly confidential documents, it focuses on developing AI models and its own RAG architecture.

RIKEN Suuri Co., Ltd.

Head office: Chiyoda-ku, Tokyo / President & CEO: Tetsuya Eda / https://www.riken-suuri.jp

RIKEN Suuri was established on October 1, 2020, through joint investment by RIKEN, RIKEN Innovation Co., Ltd., and JSOL Corporation. By coordinating a co-creation platform between industry and academia, it delivers the DX value of the future.

JSOL Corporation

Head office: Chiyoda-ku, Tokyo / President & CEO: Takeshi Nagai / https://www.jsol.co.jp

JSOL Corporation is a system consulting and solution integrator funded by NTT DATA Corporation and The Japan Research Institute. Since its founding in 2006, it has drawn on deep business know-how and system development capabilities cultivated in manufacturing, distribution and services, finance, and the public sector, along with advanced analysis technology in engineering science, to contribute to raising customer value.

Contact


※ The contents described in this release are subject to change without notice. ※ Product names and company names in this release are trademarks or registered trademarks of their respective companies.