ナンバーワンソリューションズ List of Products and Services

"Microcosm" is a local AI system that can be utilized with systems owned and operated in-house. Since it operates in an offline environment, it allows for document creation and data analysis while protecting confidential information and customer personal data. You can easily implement the necessary LLM and customize it according to your work needs. Any LLM that is publicly available as open source can be used. [Features] ■ Usable without the internet ■ Ability to choose suitable LLMs ■ Free to use ■ Prompt and RAG settings available *For more details, please download the PDF or feel free to contact us.

This document introduces how local LLMs are utilized across various industries and the results companies are achieving, based on specific examples. It includes the four benefits and use cases of local LLMs, technical requirements, challenges in implementation, and successful cases in sectors such as finance and healthcare. Local LLMs significantly contribute to data privacy protection, cost reduction, and operational efficiency. [Contents (partial)] ■ Introduction ■ Four Benefits of Local LLMs ■ Use Cases of Local LLMs ■ Technical Requirements of Local LLMs ■ Challenges in Implementation *For more details, please download the PDF or feel free to contact us.

This document summarizes efforts to utilize generative AI to convert the tacit knowledge of skilled technicians into explicit knowledge, addressing the challenges of technology transfer. It discusses the overall challenges in the industry, the factors contributing to these challenges, solutions, successful examples of AI search utilization, the benefits of implementation, and local AI applications that can use AI search. By leveraging AI search powered by generative AI technology, it is possible to find the desired data from personalized information and siloed internal data. 【Contents】 ■ Introduction ■ Challenges in the Industry ■ Examples and Effects of Generative AI ■ Benefits of Implementing AI Search ■ What are Local AI Applications that Can Use AI Search? *For more details, please download the PDF or feel free to contact us.

This document introduces how local LLMs are utilized across various industries and the results companies are achieving, based on specific examples. It includes conditions for high-security environments, types of environments, and the benefits of local LLMs, as well as industry-specific success stories of local LLM applications in sectors such as finance and healthcare. Generative AI in high-security environments balances security and efficiency. Let's start transforming your business with your local LLM. 【Contents】 ■ Introduction ■ Why is security important for generative AI? ■ Industry-specific success stories of local LLM applications ■ Conclusion ■ What are applications that make local LLMs easy to use? *For more details, please download the PDF or feel free to contact us.

We will introduce a case study of AI search utilization by another company. Demonstration experiment of the railway construction skill inheritance system at JR West Japan. As background for the project, there was a situation where the formalization of tacit knowledge was not progressing, and young engineers were unable to quickly search for the necessary information. Through the demonstration experiment, young engineers were able to quickly obtain the necessary information, and a new method for quantifying the tacit knowledge of veterans promoted skill inheritance. [Case Overview] ■ Background - Aging of skilled construction personnel is progressing. - Approximately 60,000 construction plans and accident reports are stored in a distributed manner on the cloud. ■ Results - The accuracy of discovering inconsistencies in construction plans improved by 45%. - The relevance score of search results improved to 0.82 (1.4 times compared to before). *For more details, please download the PDF or feel free to contact us.

Introducing examples of AI search utilization from other companies. Taisei Corporation's Construction Technology Exploration System The company experiences over 3,000 inquiries regarding construction technology each year. On average, it takes 3 business days to respond, and the barrier of technical jargon hinders the autonomous learning of young engineers. After implementation, there was a significant reduction in inquiry response time and efficient technology transfer through continuous improvement. [Case Overview] ■ Background - Inefficiency of an OJT-dependent education system ■ Results - Immediate resolution rate improved from 58% to 89% - Time to search for similar cases reduced from an average of 45 minutes to within 5 minutes - Response accuracy improved by 2.3% each month due to feedback loops *For more details, please download the PDF or feel free to contact us.

We will introduce examples of AI search utilization from other companies. Inheritance of product development know-how at a chemical manufacturer (Lion Corporation × NTT Data) The company faced challenges such as the inability to inherit tacit knowledge, including difficult-to-quantify raw material blending ratios and foaming control techniques. After implementation, they achieved maintenance of product development quality, increased productivity, and more efficient know-how inheritance. 【Case Overview】 ■ Background - The turnover rate of skilled technicians in the development of laundry powder detergent manufacturing processes is 7% per year. ■ Results - The average technical acquisition period was reduced from 6.8 months to 3.2 months. - The first-pass yield in the optimization process of detergent particle size improved from 72% to 89%. - The time spent on knowledge extraction tasks by skilled technicians was reduced by 58%. *For more details, please download the PDF or feel free to contact us.

We will introduce a case study of AI search utilization from another company. Danger Prediction AI from a chemical materials manufacturer (Asahi Kasei) The company faced challenges such as the limitations of a safety management system that relied on the danger prediction abilities of experienced workers in chemical plants. After implementation, they achieved a strengthened safety management system and improved work efficiency by 1,200 hours annually. 【Case Overview】 ■ Background - A 34% turnover rate among veteran workers over the past five years, with a sharp increase in near-miss incidents. ■ Results - The number of serious accidents decreased by 67% compared to the previous year. - Danger prediction judgments were reproduced with 92% accuracy. - The risk assessment time for reaction tank cleaning tasks was reduced from 45 minutes to 8 minutes. *For more details, please download the PDF or feel free to contact us.

We will introduce case studies of AI search utilization from other companies. Design Knowledge Inheritance Platform at NEC The company faced challenges such as 3D-CAD data in the manufacturing industry being isolated within the organization and past design know-how not being utilized. After implementation, the proposal for optimal placement of welding points extended the fatigue life to 2.3 times compared to before. This promoted the efficiency of the design process, cost reduction, and effective utilization of design know-how. [Case Overview] ■ Background - A heavy electrical manufacturer had a reuse rate of only 17% for 30 years of design data. ■ Results - The average number of prototypes decreased from 4.2 to 1.8. - The man-hours for design changes were reduced by 37%. - The time for inheriting vibration design knowledge was shortened by 80%. *For more details, please download the PDF or feel free to contact us.

We would like to introduce the "Benefits of Implementing AI Search" that we propose. By utilizing generative AI to convert tacit knowledge into explicit knowledge, the education of young engineers is made more efficient, reducing learning time. Centralized information management allows for quicker problem-solving on-site, and it may significantly reduce the risk of technology leakage. 【Benefits】 ■ Improved operational efficiency: Rapid problem-solving through centralized information management ■ Enhanced educational efficiency: Shortened learning time through the formalization of tacit knowledge ■ Risk reduction: Prevention of technology leakage and sustainable technology succession *For more details, please download the PDF or feel free to contact us.

"Microcosm" is a local AI application that allows anyone to easily and safely utilize AI, enabling quick searches for desired files through AI search. It can operate in an offline environment, allowing for document creation and data analysis while protecting confidential information and customer personal data. Since prompts and RAG settings can be configured in advance, suitable results can be generated with just text input. By utilizing RAG, FAQ chats can be easily created. 【Features】 ■ Usable without internet ■ Choose and use suitable LLMs ■ No-code, ready to use immediately ■ Prompt and RAG settings available *For more details, please download the PDF or feel free to contact us.

I would like to introduce the reasons why security is important for generative AI. Due to the increased risk of data leaks that can lead to loss of customer trust and legal liability, as well as the rise in cyberattacks such as hacking and ransomware targeting AI systems, it is crucial to operate generative AI in a highly secure environment. Local LLMs minimize the risk of leaks by not requiring external transmission and offer benefits such as reducing the burden of document creation and automating data analysis. [Conditions for a Highly Secure Environment] ■ Data Sovereignty: Data storage and processing are completed within the company's environment. ■ Access Control: Thorough authentication processes and authority management. ■ Encryption: End-to-end encryption of communication and stored data. ■ Monitoring and Auditing: Log management and early detection of unauthorized access. *For more details, please download the PDF or feel free to contact us.

We would like to introduce a case study on the use of local LLMs at JP Morgan Chase (financial industry). The company utilizes its own developed local LLM, "LLM Suite," which is used by approximately 50,000 employees in the asset management division. Key points of implementation include customer data management under strict regulatory environments, as well as providing document creation, idea generation, and summarization functions. 【Case Overview】 ■ Key Points of Implementation - Customer data management under strict regulatory environments - Provision of document creation, idea generation, and summarization functions ■ Best Practices - Model design focused on regulatory compliance - Implementation of training programs for employees *For more details, please download the PDF or feel free to contact us.

We will introduce a case study of local LLM utilization by another company. Oda Hospital (Healthcare Industry) The hospital utilizes the local LLM "OPTiM AI" in conjunction with electronic medical records, operating in a completely local environment. The key points of implementation are the automatic generation of admission and discharge summaries and the improvement of accuracy through customization of medical terminology. 【Case Overview】 ■ Key Points of Implementation - Automatic generation of admission and discharge summaries - Improvement of accuracy through customization of medical terminology ■ Best Practices - Model adjustments that balance safety and efficiency - Close collaboration with the IT department *For more details, please download the PDF or feel free to contact us.

We will introduce a case study of local LLM utilization by another company. Mie University Hospital (Medical Industry) The hospital operates a discharge summary creation process using NTT's "tsuzumi" within its internal network. The key points of the implementation are the improvement of electronic medical record summarization accuracy and processing patient data without external transmission. 【Case Overview】 ■ Key Points of Implementation - Improvement of electronic medical record summarization accuracy - Processing patient data without external transmission ■ Best Practices - Optimized for medical environments with a small-scale model - Stabilizing the process through phased implementation *For more details, please download the PDF or feel free to contact us.

We will introduce a case study of local LLM utilization by another company. Tsuchiya Gosei Co., Ltd. (Manufacturing Industry) The company faced challenges with data management and system optimization. Therefore, they achieved advanced automation through LLM tuning specialized for the manufacturing process. Key points of implementation include real-time data analysis and automation of quality control, as well as the acceleration of problem detection and response. 【Case Overview】 ■ Key Points of Implementation - Automation of real-time data analysis and quality control - Acceleration of problem detection and response ■ Best Practices - Model customization corresponding to each process - Strengthening collaboration with on-site personnel *For more details, please download the PDF or feel free to contact us.

We would like to introduce a case study on the use of local LLMs at JP Morgan Chase (financial industry). The company utilizes its own developed local LLM, "LLM Suite," which is used by approximately 50,000 employees in the asset management department. The key points of implementation are customer data management under a strict regulatory environment, as well as providing document creation, idea generation, and summarization functions. 【Case Overview】 ■ Key Points of Implementation - Customer data management under a strict regulatory environment - Provision of document creation, idea generation, and summarization functions ■ Best Practices - Model design focused on regulatory compliance - Implementation of training programs for employees *For more details, please download the PDF or feel free to contact us.

The document discusses "Five Challenges That Can Be Solved with Generative AI Enterprise Search." It introduces issues such as the enormous amount of time employees spend searching for information, which decreases operational efficiency, and the challenge of cross-system searches due to internal information being dispersed across multiple systems. Additionally, it explains the challenges faced when implementing local LLMs and their solutions. Please take a look. 【Contents】 ■ Introduction ■ Five Challenges That Can Be Solved with Generative AI Enterprise Search ■ For Those Who Want to Utilize Generative AI Enterprise Search While Protecting Confidential Information ■ Conclusion *For more details, please download the PDF or feel free to contact us.

We will introduce the effectiveness of implementing "Generative AI Enterprise Search" in addressing the issue of wasting a vast amount of time on information gathering. Employees spend a significant amount of time each day searching for necessary information. It is said that, on average, they dedicate over 17.8 hours a week to information gathering, which leads to a decrease in operational efficiency. With the introduction of this system, a reduction of approximately 1.2 hours in work time per week is expected. 【Case Overview】 ■Challenges - Excessive information gathering time for employees (over 17.8 hours a week) - Decreased operational efficiency, hindering focus on core tasks ■Implementation Effects - There are cases where the Leena.ai platform achieved information access that is 90% faster than general searches, resulting in over three times the productivity for teams. *For more details, please download the PDF or feel free to contact us.

We will introduce the effectiveness of implementing "Generative AI Enterprise Search" in addressing the issue of insufficient accuracy in search results. General keyword searches rely on matching the words input by the user, which often fails to adequately reflect the search intent and context. As a result, users frequently struggle to find the information they truly seek, which can hinder decision-making and responses in business operations. AI search engines have a higher contextual understanding capability than typical search engines, allowing them to provide more relevant results. [Case Overview] ■Challenges - Limitations of keyword matching: Lack of understanding of context and background - Difficulty in obtaining information that aligns with the user's true needs ■Implementation Effects - In the iFinder software, it has been confirmed that experienced employees can search for information 11 times faster through the service hotline. *For more details, please download the PDF or feel free to contact us.

We will introduce the effects of implementing "Generative AI Enterprise Search" in response to the challenge of integrating distributed data sources. Internal information is managed across multiple systems such as email, cloud storage, CRM, and ERP, making it difficult to search across them. As a result, necessary information gets buried, hindering rapid decision-making in business operations. By integrating multiple data sources and providing a unified search environment, we can prevent information loss. 【Case Overview】 ■Challenges - Information dispersion across multiple systems is a significant barrier. - Conducting cross-system searches is difficult, leading to information loss. ■Implementation Effects - With Google Vertex AI Search, the indexing time for small datasets has been reduced from several hours to just a few minutes, achieving improved processing speed. *For more details, please download the PDF or feel free to contact us.

We will introduce the effectiveness of implementing "Generative AI Enterprise Search" in addressing the challenge of multilingual information management. In global companies, documents created in different languages exist, leading to language barriers in information retrieval and sharing, which poses a risk of decreased communication efficiency. GenAI can automatically provide context-appropriate translations in real-time, significantly reducing translation costs and time. [Case Overview] ■ Challenges - Complicated information retrieval between different languages - Need for summarization and translation of multilingual documents ■ Implementation Effects - Flexibly accommodates the addition of new languages and increased communication volume as the organization expands *For more details, please download the PDF or feel free to contact us.

We will introduce the effectiveness of implementing "Generative AI Enterprise Search" in addressing concerns related to information security and privacy. The risks of information leakage and unauthorized access associated with AI are concerns faced by many companies. Strict security measures are particularly required when handling confidential information. With ACL (Access Control List), detailed permission settings can be configured based on roles and departments. 【Case Overview】 ■Challenges - Concerns about information leakage and unauthorized access - Access rights management and authentication functions are essential for protecting confidential information ■Implementation Effects - Google Vertex AI Search uses Customer Managed Encryption Keys (CMEK) and meets high security standards such as HIPAA and the ISO 27000 series. *For more details, please download the PDF or feel free to contact us.

"Microcosm" is a local AI application that builds a dedicated AI for your company using local LLMs and semantic databases. It allows for document creation and data analysis while protecting confidential information and customer personal data, as it can operate in an offline environment. Additionally, once installed on a PC, it can be used immediately without coding, and there are no token costs. We will support the necessary environment setup for implementation within departments or across the entire company. 【Features】 ■ Usable without internet ■ Usable immediately without coding ■ Ability to choose the appropriate LLM ■ Prompt and RAG settings available *For more details, please download the PDF or feel free to contact us.

We would like to introduce a case study on the use of AI search in the construction technology exploration system of Taisei Corporation. The company receives over 3,000 inquiries related to construction technology each year. On average, it takes 3 business days to respond, and the barrier of technical jargon has hindered the autonomous learning of young engineers. After implementation, there was a significant reduction in response time for inquiries and efficient technology transfer through continuous improvement. 【Case Overview】 ■ Background - Inefficiency of an OJT-dependent education system ■ Results - Immediate resolution rate improved from 58% to 89% - Time to search for similar cases reduced from an average of 45 minutes to under 5 minutes - Response accuracy improved by 2.3% each month due to a feedback loop *For more details, please download the PDF or feel free to contact us.