The automotive industry faces a critical challenge in enhancing product value through electrification and intelligence, requiring additional efforts beyond the standard value of traditional vehicles. Efficient resource use will directly impact product value and business sustainability. In today's fast-paced world, understanding evolving customer needs is crucial for timely product delivery. One initiative involved implementing an Advanced Expert System (A-ES) to transfer skilled engineers’ knowledge to younger ones, starting with the review process for collision-safety vehicle development. Crash simulation work, crucial for this process, often took more than a day per simulation, with errors requiring major rework. A-ES streamlined simple tasks, freeing up time for value creation. However, modeling know-how for A-ES was time-consuming, taking 400 hours to create a knowledge model for just two to three components out of over 20,000 car parts, posing a challenge for wider business deployment.
To enhance knowledge modeling efficiency, IBM proposed using generative AI to extract and database knowledge from Microsoft PowerPoint materials containing dispersed know-how within the company. Honda’s skilled technicians have left valuable knowledge in PowerPoint documents, which are rich in diagrams and graphs but sparse in text, making AI-based reuse challenging. IBM suggested applying a large, multi-modal model (LMM) to convert graph and diagram content into text, improving the reuse of knowledge in AI-rich PowerPoint materials. Storing textual know-how in a database allows for retrieval-augmented generation (RAG) search-like knowledge utilization. An IBM® watsonx.ai™ pilot conducted from November to December of 2023 validated the feasibility of this approach.
With conventional A-ES, an experienced skilled engineer would take three years to create a handbook and one year to create a model from the handbook. The man-hour savings achieved through A-ES were 30% for development and 50% for planning/management. Using generative AI, Honda’s technical documentation can now be modeled as sentences, reducing modeling time from three years to one. This approach expands document utilization areas and boosts business efficiency. IBM demonstrated full coverage from value validation to delivery and operation, along with watsonx.ai’s infrastructure model concept. By applying LMM and LLM, IBM ensured project feasibility, leading to proof of concept projects and future production development activities.
Honda (link resides outside of ibm.com) is a Japanese public multinational conglomerate manufacturer of automobiles, motorcycles, and battery-powered equipment, headquartered in Minato, Tokyo, Japan. Since 1959, it's been the top motorcycle maker, producing 400 million by 2019. It's also the biggest internal combustion engine maker, with over 14 million engines yearly. Honda became Japan's second-largest car maker in 2001 and was the eighth largest globally in 2015.
© Copyright IBM Corporation 2024. IBM, the IBM logo, and watsonx.ai are trademarks or registered trademarks of IBM Corp., in the U.S. and/or other countries. This document is current as of the initial date of publication and may be changed by IBM at any time. Not all offerings are available in every country in which IBM operates.
Client examples are presented as illustrations of how those clients have used IBM products and the results they may have achieved. Actual performance, cost, savings or other results in other operating environments may vary.