Spark+ and Somic Ishikawa Embark on Joint Development of AI for Automatic Generation of DRBFM Worksheets to Reduce "Burden on Design Sites"

SparkPlus Inc. (Headquarters: Bunkyo-ku, Tokyo; President and CEO: Jumpei Honda; hereinafter referred to as "Spark+") and Somic Ishikawa Inc. (Office: Hamamatsu City, Shizuoka Prefecture; President and CEO: Kaname Saito; hereinafter referred to as "Somic Ishikawa") are pleased to announce their joint development of an AI agent for automatic generation of DRBFM worksheets, aimed at improving the efficiency of design reviews in the manufacturing industry.

Since its inception, Spark+ has operated as a solutions company "tackling industrial challenges with AI," providing comprehensive support from the development to the implementation of AI solutions across a wide range of fields, including manufacturing. In the design and development domain of manufacturing, in particular, Spark+ has been engaged in research and development of AI algorithms specialized for design since its establishment, with a proven track record in advancing the utilization of design information, technical documents, and past knowledge.

In this joint development, Somic Ishikawa, which holds the top domestic market share for "ball joints," automotive suspension components, will combine its accumulated design knowledge and technical expertise with Spark+'s proprietary AI algorithms and utilization knowledge in the design domain to develop an AI agent that supports the creation of DRBFM worksheets.

Background of the Initiative

In the design and development sites of the manufacturing industry, the amount of information handled by designers is increasing due to product sophistication and shortening development periods. To appropriately identify the impacts and risks associated with design changes and ensure quality, it is crucial to enhance design reviews by cross-referencing past defect information, design review history, evaluation results, and technical documents.

DRBFM (Design Review Based on Failure Mode) is an important method for improving design quality by focusing on design changes and organizing potential failure modes, their impacts, causes, and countermeasures. However, the creation process requires designers to create worksheets while reviewing vast amounts of documentation, consuming significant effort.

Specifically, the following points are areas where many designers feel challenges in efficiency:

- The enormous effort required for creating documentation associated with DRBFM worksheet creation. - The effort involved in searching and confirming past design information, defect information, and technical knowledge. - The transfer of tacit knowledge and judgment perspectives held by experienced designers to younger designers. - Standardization of design review quality and reduction of reliance on individual expertise.

Spark+ has previously engaged in AI utilization for DRBFM creation tasks through joint development with automotive parts manufacturers, demonstrating significant effort reduction effects in some work processes. In this joint development, Spark+ will deploy its accumulated expertise to Somic Ishikawa's design domain.

Against this backdrop, Spark+ and Somic Ishikawa will work to build a system that supports DRBFM creation tasks using AI, while leveraging the expertise of designers.

Overview of This Joint Development

This joint development will combine Somic Ishikawa's design knowledge cultivated through the development, design, and manufacturing of automotive parts with Spark+'s design-domain-specific AI algorithms, utilizing generative AI and multimodal AI technologies. Through this, we will proceed with the development of an AI agent that supports the automatic generation of DRBFM worksheets.

The AI agent to be developed aims to organize the perspectives to be considered in DRBFM based on design change information and related documents, and to present potential failure modes, causes of occurrence, impacts, verification items, and proposed countermeasures in a worksheet format.

Furthermore, by not only generating text but also referencing past technical knowledge and similar cases to present perspectives that designers should confirm during reviews, we aim to achieve both improved efficiency and enhanced quality of design reviews.

Future Direction

Moving forward, we will aim to build an AI agent that can be utilized in practical work by progressively organizing target tasks and reference data, designing evaluation of generated items, conducting reviews by designers, and improving output accuracy. Spark+ will further promote AI utilization in the design and development domain of manufacturing and continue to support Somic Ishikawa's design review efficiency and advanced utilization of technical knowledge.

Comment from Mr. Kazumasa Takahashi, General Manager of Design Department, Design General Affairs Office, Somic Ishikawa Inc.

"DRBFM is the core of design quality, but extracting information from vast related documents and inputting it into worksheets has been a significant burden for designers. Through this joint development, we expect a substantial reduction in creation effort as AI will automatically organize and assist in inputting past knowledge and defect examples. By streamlining simple input tasks, we aim to build a system that allows designers to allocate sufficient time to the original objective of 'deep consideration of risks,' thereby further improving product quality."

Comment from Mr. Jumpei Honda, President and CEO, SparkPlus Inc.

"DRBFM is a crucial process for structurally examining risks associated with design changes, and it is an area with significant potential for AI-driven support. We are extremely pleased to embark on the joint development of an AI agent to support DRBFM worksheet creation with Somic Ishikawa, who have honed their technology over many years in the design and manufacturing of automotive parts. Through this joint development, we will contribute to the efficiency and quality improvement of design review operations, and further to the advancement of technology transfer in the manufacturing industry."

Company Overview of Somic Ishikawa Inc.

Company Name: Somic Ishikawa Inc. Representative: President and CEO, Kaname Saito Location: 500 Furukawacho, Chuo-ku, Hamamatsu City, Shizuoka Prefecture Founded: May 1916 Business Activities: Development, design, manufacturing, and sales of automotive parts (ball joints) URL: https://www.somic.co.jp/

Company Overview of SparkPlus Inc.

Spark+ is a startup spun out of the University of Tokyo's Matsuo Laboratory, providing solutions that leverage cutting-edge AI technology. In approximately two years since its founding, it has implemented AI-driven DX services through collaborative projects with numerous major companies, including Mitsubishi Heavy Industries, Toyota Industries, and Lifenet Insurance.

Company Name: SparkPlus Inc. (Commonly known as: Spark+)

Representative: President and CEO, Jumpei Honda

Location: Hongo Tsuruna Bldg. 6F, 6-17-9 Hongo, Bunkyo-ku, Tokyo

Business Activities: AI solutions and consulting for industrial applications

Established: February 2024

URL: https://sparkplus.co.jp

【Contact Information】

SparkPlus Inc. Public Relations

Email: [email protected]