How Generative AI Helped Final Aim to Design a Unique Yamaha EV for Farming Terrain

YASUHIDE YOKOI: Hello. My name is Yasuhide Yokoi, co-founder and chief design officer at Final Aim. Today, I'd like to talk about our latest design collaboration with Yamaha Motor by the title, "How Generative AI Helped Final Aim to Design a Unique Yamaha EV Platform for Farming Terrain."

First, I would like to introduce myself and my company. Once again, my name is Yasuhide Yokoi, co-founder and chief design officer at Final Aim. Please call me Yasu. Born in Japan and raised in Australia, I started an industrial designer career at Nikon Corporation, designing mass produced products such as digital cameras, binoculars, and microscopes. After seven years of experience, I joined a 3D printing platform startup, Kabuku, as an initial founding member. After the startup's M&A, I co-founded my current startup Final Aim.

Over my 15 plus years of industrial designer career, I have collaborated with numerous enterprises, such as Toyota, Honda, Sony, Panasonic, Olympus, SoftBank, Microsoft and Autodesk. In my previous startup, for example, I collaborated with Toyota as a project leader and industrial designer to establish Toyota's first 3D printed mass customization service. The service lets users customize the car parts online. After the customization, the 3D data is sent through the 3D printing platform, which is connected to factories around the globe. The original parts were delivered to users within two weeks. The service not only created new business to Toyota, but also strengthened the relationship between Toyota brand and its customers.

As another example, with Honda, I developed a 3D printing single-seater electric vehicle. The car was fully customized for brand, Toshimaya, which is the nationally famous sweets company in Japan. The sweets company was seeking a fully original delivery car that would be loved by the locals and represent its strong brand image for promotion. Utilizing large industrial 3D printers around the world, the project was able to provide an original design within a couple of weeks, quickly meeting the local needs.

Autodesk Fusion was fully utilized in the design development process to meet the quality and speed. As a side note, I have been an Autodesk Fusion user for over 10 years. And I'm probably one of the first users in Japan. My journey with Autodesk Fusion has been exciting. And I'm very impressed by the aggressive evolution over the years, integrating all kinds of design and manufacturing technologies, including 3D printing and AI.

Recently, I have collaborated with a Japanese Katana swords craftsman, who inherited more than 400 years of crafting history and combined the latest approach of generative design by Autodesk Fusion. By leveraging generative design, the product achieved functionality, such as grip, ergonomics, strength, material selection, and less consumption. Furthermore, not just its functional achievements, this innovative approach generated a breathtaking, organic look, adding an energetic impression to the sword. As you know by now, my lifetime passion is to make design and manufacturing a better industry by leveraging the latest technologies. My exploration of advance design and manufacturing exciting now more than ever.

Let me introduce my current company-- Final Aim. With headquarters in the US and a subsidiary in Japan, Final Aim's a startup, driving the global adoption of smart contract technologies in design and manufacturing. With its mission to unlock the value of design and manufacturing, the company has collaborated with Sony Design, Panasonic headquarters in New York, and numerous international companies. The startup was accepted into Berkeley SkyDeck, a world renowned Silicon Valley accelerator, and was chosen the most likely to become the next unicorn. Final Aim also won the Japan-US Innovation Award, a co-program of the Japan Society of Northern California and Stanford University.

With my industrial design capability, which was provided to Yamaha Motor, the company also develops and provides its design and IP management platform. We call it Final Design. It is a single source platform for all essential data, from design files to legal contracts and documents. Final Design ensures secure, tamper proof data storage with legal compliance, supervised by top international law firms, addressing evolving intellectual property risks in generative AI. This platform was also the core of the collaboration with Yamaha Motor. I will explain more in detail later in this presentation.

Now, I'd like to go over the session overview. In this session, I will mention that Yamaha Motor and Final Aim have collaborated on a new EV prototype, "Concept 451." The Concept 451 project creates electric mobility using a small EV platform, specialized for low-speed personal mobility and featuring a unique design for light work in agricultural areas. I will explain how generative AI is becoming a major force in design heavy industries, improving the communication between clients and designers.

As learning objectives, you will be able to understand how generative AI can accelerate the design process by reducing iteration time and increasing creativity. You will learn how cross-disciplinary collaboration and AI-driven visuals enhance decision making and team alignment. Furthermore, you'll be able to understand the importance of securing intellectual property and supporting innovation in the design and manufacturing process.

And below are the key takeaways for this session. First, generative AI can quickly generate thousands of design concepts, significantly reducing design cycle times. Second, utilizing AI improves the likelihood of design adoption leading to a higher success rate. And lastly, Final Aim's design and IP management platform, Final Design, ensure secure and authentic design to protect innovation and streamline collaboration with tools like Autodesk Fusion.

Now, I'd like to share the project overview. I will share how the project was formalized and proceeded from concept, design, and manufacturing. Firstly, let me share the project's background and how Yamaha Motor initiated it.

Before this mobility project, Yamaha Motor's New Business Development Division launched a new EV platform that can be designed flexibly. With this platform, Yamaha Motor is exploring a wide range of application ideas across various fields and seeking co-creation and collaboration partners. A key feature is the openness to collaboration with co-creation partners, regardless of size or area of expertise. Specifically, based on this platform, they sought to explore the possible mobility that solves issues surrounding the agriculture industry. Design should be functional and suggest a new direction for the agriculture industry.

Furthermore, the team at Yamaha Motor had hypotheses utilizing generative AI and cloud 3D software, like Fusion, to realize the agile approach in new business development. In new business ventures, the ability to move design and development forward in an agile way and test it in the market is very crucial. Another strong intention that drove the project forward aggressively was that the design prototype would debut at Tokyo Auto Salon.

It is one of the largest motor shows in Japan with more than 230,000 attendees and international media exposure. And it was only three months away from the project kickoff-- very scary. This strict timeline also pushed the team to integrate new tools like generative AI and Autodesk Fusion.

With this backgrounds, big and small, we had numerous challenges to overcome. On the highest level, as you can see on the slide, we had to balance three main factors-- quality, efficiency and security. First, quality, ensuring creativity and collaboration remain at the forefront, even as we integrated new technologies like generative AI. Second, efficiency, being able to speed up the process without sacrificing the depth of thoughtfulness of the design. And finally, security, protecting the intellectual property of the designs, especially when using the generative AI, which introduces new consideration in terms of data ownership and traceability. These challenges were at the core of approach, and overcoming them was crucial to the success of this project.

Now, I would like to describe the process overview. From the kickoff from zero concept, we utilize generative AI tools as much as possible. To complete the project in just a few months, we conducted a process trial to see how much generative AI could be applied across the workflow. This workflow covered three phases-- concept and planning, industrial design, production and launch. As I will explain in detail later, in addition to promoting the usage of generative AI, the team also stressed ensuring and managing intellectual property and authenticity throughout the total process.

And before explaining each process in depth, I would like to go over some basics about generative AI. It may be too basic for some of you, but it will be essential to understand beforehand, especially the different input and output the generative AI tools provide. The first type, as you see in the slide, is text-to-text AI. As you may have used or at least tried once, it is one of the foundational categories of generative AI. It works by taking the input of text, or in other words, a prompt, and generating new written content based on that.

This type of AI is especially useful in early stage brainstorming, drafting, planning, or when you need a quick way to explore new ideas. It helps to speed up the learning process. But more importantly, it opens up new possibilities by offering unique perspective and suggestions you might not ever think of.

Text-to-image AI tools are another exciting category of generative AI. Instead of generating texts, these tools convert written input into visual content. You can provide a description, like autonomous mobility, futuristic, chrome exterior, sleek canopy, et cetera. And the AI will generate multiple images that match that description.

This is incredibly useful during the early design stages when you're trying to visualize concepts quickly. It allows designers to explore a wide range of visual ideas quickly, offering inspiration or even helping designers find directions you hadn't thought of before. It's a powerful way to bridge the gap between abstract and visual representation.

Text-to-image is especially valuable for prototyping. It speeds up turning ideas into reality, allowing teams to iterate faster and making informed decisions based on the visual generated. It makes the creative process more dynamic and accessible, allowing teams to focus on refining their vision.

And lastly, the third type, image-to-image AI tools are fascinating part of generative AI. They take an existing image as an input and transform it into new variations or improved versions. For example, you can input a basic hand-drawn sketch or low resolution design, and then the AI will enhance it. You can further optimize it with additional inputs and parameters, like text prompts, to generate alternative styles or modify the visual elements based on the specific parameters. This, too, is incredibly powerful for designers and engineers to refine ideas during design and development.

Let's say you have an early stage design concept for a product. You can use image-to-image to explore different color schemes, materials, or even shapes without starting from scratch each time. Image-to-image AI is particularly valuable for designers of product development or architecture, fashion or even marketing materials. It allows teams to take on an existing concept quickly and experiment with variations, leading to faster decision making and more refined results. In short, it's a powerful tool for evolving and refining designs quickly and efficiently. And now, AI has evolved even further with capability to generate movies and 3D models, opening up new frontiers in design.

Now that we've discussed the overall approach and goals of this project, let's dive into how the team actually used AI to drive the design process. In the initial concept and planning phase, just like a human mentor or advisor, the generative AI tool played a key role in helping the team explore various important aspects. We used AI to investigate future environmental and societal contexts and agriculture industry challenges.

For example, in defining target users and demographics as non-industry related users, we quickly caught up on underlying problems of age population and how serious those problems are for sustainable future. We also explored the functional requirements for single-seater electric vehicle, specifically designed for use in farmland. The AI provide insights into how we could address practical challenges and improve vehicle's functionality in these unique environments. Lastly, AI helped us examine exterior design requirements, both functional and emotional aspects. We wanted the exterior design to be practical and emotionally appealing to target users, particularly to the younger generation.

In the industrial design phase, we utilize various image-based generative AI tools to support the design process. The tools range from text-to-image, image-to-image or even image-to-text, allowing us to explore different creative methods. We set detailed parameters and continually refined the prompts to ensure AI-generated outputs aligned with our design goals.

By reiterating and optimizing the input, we got more targeted and relevant results, helping us to move forward faster and better. Additionally, we compared outputs from multiple AI tools using the same input. This allowed us to analyze difference in results and select the most suitable tools designs for further refinement.

What was really interesting, though, was the unintentional design that the AI generated. Some of these results were unexpected. But they pushed us to think outside the box. These unexpected ideas help us explore directions we might not have considered otherwise, sparking new creativity. Ultimately, based on the project's intent and conditions, the AI generated over 2,000 design proposals in total. This iterative process helped us identify key features and develop a distinctive design that met both functional and aesthetic requirements.

And here, I have the designs from the final convergent phase of the project, of course, created by AI. As you can see, we explored a variety of forms and aesthetics and keeping in mind the functionality and the unique needs of agricultural environments. These designs represents the culmination of the iterative process we went through with the help of AI.

In this phase, we also focused heavily on ensuring the design maintains structural integrity while providing the flexibility needed for different farming conditions. By integrating Yamaha Motor's existing expertise in vehicle development and with my intuition as an industrial designer, we refined and narrowed down the design to one idea. I believe this is an area where human experience and knowledge are still essential and something AI can't fully replace.

In parallel with the generative AI ideation, we utilize Autodesk Fusion in the early stages of the process. Fusion played a critical role in ensuring seamless data integration, especially with Yamaha Motor's EV platform. This allowed for efficient communication and collaboration between our design, engineering and manufacturing teams. Fusion's ability to integrate across various stages helped us streamline the workflow. It was particularly important in this project where time was tight and precision was essential.

Furthermore, our team was spread across multiple locations, me in the US, Yamaha team in Tokyo, Hamamatsu and Kyoto in Japan. So Fusion's cloud functionality became a critical role for maintaining communication and collaboration. We could seamlessly share design data in real time, enabling faster iterations and feedback loops. This was key to keeping everyone aligned, ensuring we hit all milestones, even with the rapid pace of development. By leveraging Fusion's capability, we made a challenging timeline while maintaining a high quality in the final mobility design.

And here's the final production model. After a very short timeline of collaboration and leveraging both AI and Autodesk Fusion, we arrived at this design. Once again, about the Yamaha Concept 451 is a micro agricultural EV based on Yamaha Motor EV platform. It addresses many problems that lie behind the agriculture industry, for example, the aging population. We have designed both functional and emotional aspects to make it easier for younger generations who is not related to agriculture industry yet to engage in this industry. The model embodies both practicality and innovation, reflecting the essence of the project goal.

And this is another angle of the vehicle. And this is another angle from the rear of the vehicle.

And here's the moment of the public launch at the Tokyo Auto Salon. The response was overwhelmingly positive, with attendees excited about the design and the concept behind it. The event gave us a chance to showcase the final model and connect with a wider audience, generating buzz for this innovative take on agricultural vehicles.

And specifically, we received a lot of positive feedback and comments from younger people, saying kawaii, meaning it's a cute design in Japanese. This is something no other agriculture mobility has. And it represents a new innovative design that we are able to achieve.

Now, I'd like to talk about the results and what were the achievements and successes of this project. As I mentioned in the key challenges, we focused on three key achievements throughout this project-- quality, efficiency, and security. First, quality was driven by creativity and collaboration across teams. Second, efficiency was achieved through the speed and quantity of AI-generated ideas. Finally, security, ensure that the intellectual property of generative AI was protected, giving us peace of mind as we pursue innovation.

One of the standout achievements in this project was a significant reduction in time for concept development. Starting from complete zero, we were able to use AI to quickly catch up on industrial knowledge and challenges. It was vital, as we were not familiar with the agriculture industry. And this allowed us to solidify the concept and planning within just a week.

By combining generative AI with the robust capability of Autodesk Fusion, we further accelerated the process. A project like this that would typically take six months to a year was completed in just three months. This dramatic reduction in development time highlights how AI and modern design tools can help teams work faster and more efficient without sacrificing the quality.

One of the most impressive outcomes of the project was the accelerated ideation process. By leveraging generative AI, we're able to generate over 2,000 design concepts in just two to three weeks. This dramatically enhanced our creativity and efficiency, allowing us to explore a wide range of possibilities in a short period of time. With such a large volume of design options, we could quickly identify key ideas and refine them into more focused concepts. The AI's ability to rapidly generate variations was crucial to our process, pushing our creative boundaries and speeding up the decision making.

And once again, generative AI played a crucial role in pushing the boundaries of our creativity throughout this project. It allowed us to explore newer design directions and even hardware details that we wouldn't have initially considered. This technology enabled us to move beyond traditional concepts, helping us approach the vehicle design from fresh and exciting perspective.

For example, AI suggested some key features that ended up being integral to the final design. One of these is the basket-like silhouette. This silhouette will give mobility a maximum luggage and customizable utility. The idea added both aesthetic value and practical functionality to the design.

Another standout feature is the pillar-less roof design. It gives a clear visibility when driving with a fresh appearance. The idea challenges conventional vehicle structures and presents an innovative way to rethink the vehicle's overall architecture.

In addition, the AI helped us come up with an asymmetrical body. This was achieved contradicting goals such as body strength, the driver protection, and the accessibility. The irregular shape gives the vehicle a distinct look that breaks away from the traditional symmetrical industrial design that we often see.

Lastly, we incorporated a striking six-lamp front face, giving the vehicle a functionality and a distinctive and bold identity. This kind of light layout and facial expression is something designers normally won't come up in the ideation. These unique elements would not have come to life without the fresh perspective of generative AI introduced. By leveraging AI as a creative tool, we are able to innovate in ways we hadn't initially imagined, matching the mobility concept and goals with a truly modern and forward-thinking design.

Another key achievement was the seamless, cross-disciplinary collaboration that generative AI enabled throughout the project. AI-driven visuals enhance communication across teams, whether it was design, engineering or marketing, allowing everyone to be on the same page more quickly. By sharing these visuals real time, we were able to align faster and make decisions more efficiently. This not only saves time, but also ensures that each discipline's unique perspective was integrated into the final design, resulting in a more cohesive and successful product.

And lastly, safeguarding intellectual property was a critical part of this project. By utilizing the design and IP managing platform, Final Design, we ensure that all design files, processes and legal documents, such as design contracts, were securely stored and verifiable. Generally speaking, intellectual property in generative AI is something designers and engineers need to take seriously.

While there are misunderstandings and fears of AI, AI is actually a tool that enhances abilities. By using it correctly and managing the design process done properly, we can use it safely and securely. And according to regulators around the world, it's essential to keep track of generative AI usage. This means that as long as I is used as creative tool and the process can be traced, the intellectual property of the design as final is fully protected. This approach not only safeguarded Yamaha's investment, but also set a new standard for responsible AI use in design, ensuring that every step of the creative process and manufacturing is transparent and verifiable.

Now let's take a closer look at our design and IP management platform, Final Design, which Yamaha Motor utilized. This platform serves as a single source for design data and intellectual property, offering secure and efficient way to manage all aspects of the project. Aside from supporting design process itself, Final Design also promotes the introduction of smart contracts into the design and manufacturing industries.

This ensures that the entire workflow from concept to production is securely tracked with full transparency and verifiability of the design, authenticity, and intellectual property. By integrating smart contracts, we're able to streamline the process, reduce risks, and enhance trust between stakeholders, making Final Design a powerful management platform for modern product development.

Final Design provides a centralized data security with a secure, traceable and tamper-proof environment for everything from design concepts to final contracts. The platform ensures full control over vital data, keeping everything protected and traceable. Teams can track the entire life cycle of information, design files, contracts, intellectual property, ensuring nothing is altered, lost or scattered.

Furthermore, the platform ensures legal compliance by being overseen and supervised by top international law firms. This guarantees that our design data management practices stay aligned with evolving legal standards, particularly those related to generative AI and intellectual property. With the platform, every file on the contracts, as we've seen in the design process earlier, is securely tracked, verified and compliant with regulations, providing proof of authenticity and legal protection. This approach gives the designers and clients, like Yamaha Motor, confidence that the data is handled in a way that meets the highest legal standards.

Lastly, Final Design is a platform that seamlessly connects with various design tools, including Autodesk Fusion. This automatic integration allows for smooth and secure design workflow, ensuring that users' intellectual property is protected at every stage of the design process. By integrating with these tools, Final Design creates a centralized hub for managing designs from concept generation to final production. This not only streamlines the creative workflow, but also ensures the integrity of IP security and the designs are maintained throughout the entire process.

To give you a closer look at how Final Design integrates with Autodesk Fusion, let's walk through the authentication process. First, we developed an original plugin for Autodesk Fusion that allows seamless data transfer. At any relevant milestone during the design process, you can simply click Send the final design on the top navigation bar and the final design is-- sorry, the final design file is automatically transferred to a platform ready for authentication. Once transferred, the data is secured using Final Design's smart contracts. This ensures that the design is protected, with all the necessary legal documents, including design contracts, legal terms, and the intellectual property information.

Finally, the platform generates a proof of authenticity, recorded in a tamper-proof format. Each record is secured with blockchain technologies, ensuring it is traceable and verifiable. As I said, the format and security are also supervised by top international law firms, ensuring full compliance with legal standards and protection.

And this is the end of my session. For those who want to revisit the story, actually, the case study is featured in Autodesk's "Design & Make." And the story is also covered in the Fusion blog from a different angle, more specifically from Autodesk Fusion. Please make sure to visit this as well.

And thank you all for taking time to join this presentation. If you have any further questions or would like to discuss the project in detail, please feel free to reach out. More than happy to discuss. Thank you.