In Episode 40 of Razorleaf’s Stay Sharp podcast, Data Science and AI: the Future of Product Development, hosts Jen Ferello and Eric Doubell, Razorleaf’s CEO, are again joined by Ramesh Haldorai, Vice President of Strategic Consulting for the 3DEXPERIENCE platform at Dassault Systèmes. This time, Jen, Eric, and Ramesh delve into topics such as how manufacturing design software has already been using AI for nearly two decades, how AI will continue to impact the manufacturing industry, and why they don’t believe AI is coming for our jobs.
The Boundaries of What We Call “AI”
To level-set, the trio discusses terminology, defining AI as systems performing tasks that typically require human intelligence. Data science takes AI a step further, using scientific methods, statistics, and computational tools—such as AI algorithms, according to Ramesh—to extract knowledge from data. To Ramesh, the key to data science is providing that knowledge in the context of the user’s work.
Eric notes that features like generative engineering have been in CAD systems for some time before AI became the next big thing. Ramesh agrees, “There’s been a lot of changes from the days that the algorithms were applied by small teams to where we are seeing AI as something that can be used by all the people in the company.” As examples, he mentions machine learning in data science—especially relevant when looking for an outcome that’s dependent on multiple parameters—as well as clustering algorithms used for activities like classifying parts into categories. “Those are ways in which we capture historical data and then use an algorithm to do something useful for the end user,” he says. “There are many places where algorithms like this are used, and sometimes the user is not even aware.”
How AI Is Changing Product Development
Ramesh describes the next level beyond generative engineering as generative experience, which he explains as the software generating multiple variations on a design and then analyzing each variation based on a basic framework of constraints the engineer provides. He adds, “If you take this concept of generating new things, you can apply it across product development. This is something that every customer should look at today and see how their existing design practices can be accelerated—and innovation can be accelerated—based on just this one capability.”
The massive computing power available today makes AI and the generative experience he described possible. It also allows for other ways to improve products, such as a closed-loop capability of comparing real-world data to predictive models. AI gives product engineers and designers the “ability to correlate between the prediction and the observed behavior. Hopefully they will learn something from this that will lead to innovative capability being delivered by the design.”
AI as an Assistant, Not a Replacement
After the trio discusses how AI can be useful in product design and how it’s likely to impact the future of the manufacturing industry, Jen poses the question that’s on so many minds: where Ramesh sees humans involved in the future of AI. While Ramesh concedes that it’s possible a large-language model AI system could generate systems engineering models for us, “because behind the scenes, systems engineering models are just math and programming,” he doesn’t see AI replacing product designers. “I don’t see AI as a capability which is equivalent to us. It is something that will help us improve our productivity,” he says. “I think it augments what we can do.”
What AI can do extremely well is act as a self-service agent, where a human, typically a consumer, asks basic questions that can be answered by AI based on observed data. But if a self-service agent is also embedded into a software or platform, the product designer can collaborate with the platform. That might be entering prompts and receiving suggestions of stored design components. It might be requesting calculations and receiving AI feedback. This is where Ramesh sees the maximum productivity gains could be found. “Until now, in our minds, engineering and manufacturing working together, that’s collaboration. But now we should open ourselves up to the possibility of collaborating with the system itself, with the AI capability. Collaborating and working together to achieve a goal—let’s say a more sustainable product.”
Learn More About Digital and Virtual Twins
The full podcast episode offers more details including the issues to watch for when incorporating AI into your product design practices, how to improve the support AI systems give you, how AI could even change consumers’ interactions with manufacturers in the future, and more.
Check out Stay Sharp Episode 40: Data Science and AI: the Future of Product Development and be sure to watch Jen and Eric’s first conversation with Ramesh in Stay Sharp Episode 38: Digital Twins vs. Virtual Twins with Ramesh Haldorai.
Join us each week for a new episode!
Follow the Razorleaf Podcast, Stay Sharp in Digital Engineering on:
