In Episode 81 of Razorleaf’s Stay Sharp podcast, COExperience 2025 Recap, co-hosts Jonathan Scott and Juliann Grant are joined by two Razorleaf colleagues to share highlights from the conference hosted by Community of Experts (COE) and Dassault Systèmes. Simulation expert Michael (Mike) Craffey works on Razorleaf’s Dassault business and 3DEXPERIENCE infrastructure and also volunteers for COE, while Presales Architect Kip Speck focuses on Dassault PDM and PLM services and applications. They recap the notable sessions and developments they experienced at the recent volunteer-run PLM industry conference.
New from Dassault Systèmes
Given that COE originated as the Catia Operators Exchange, the organization and event have a longstanding connection to Dassault, and a large part of the event centered on new products and directions for the company. A live augmented reality (AR) session, where the audience could experience the AR user’s view, was the most impressive session Mike attended. It showed, he says, “This overlap of human experience with the technology is becoming more and more commonplace.”
Another development, which Kip noted has been in discussion for at least 10 years, is Dassault’s plans to offer FedRAMP (Federal Risk and Authorization Management program) on the Dassault Systèmes private cloud, which delivers US government access to cloud services on the AWS platform, but managed and operated by US Dassault personnel. A similar expansion of capabilities is the Catia-as-a-Service offering, meaning the massive software package—which used to require commensurately large compute power—is now web-enabled and able to be run on any device anywhere, significantly expanding functionality.
As part of its seventh-generation evolution of services, Dassault also announced the new 3D UNIV+RSES (pronounced “universes”), which they describe as next-generation Dassault Systèmes technology that offers an environment for combining virtual twins, training AI agents, and protecting customer IP. Kip explains, “It’s a collection of apps and processes put together for [sectors] like industrial equipment or automotive. It’s a whole suite of roles and processes that they put together to say, here’s an industry process, things that are typical business processes and use cases for your industry.”
Kip also described a customer-perspective session that offered insights in new approaches to the 3DEXPERIENCE platform. Bell Flight discussed working with Virgin Galactic on the design of a spacecraft, using a hybrid on-premise/on-cloud approach throughout the process. “They were able to do those designs and collaborate in a better way than they’ve ever done in the past,” he says. “I thought that was a very interesting use case for two very large companies with very complex designs and engineering to collaborate.”
AI on Everyone’s Mind
According to Mike, AI “seemed to be a common theme through the whole conference. There was AI everywhere. There were more AI topics in the sessions than I’ve ever seen.” One application was predictive AI in Catia, which is gaining presence in the platform in the form of capabilities like predictive commands—similar in function to predictive text in text messaging applications. Kip notes, “Dassault has really stressed that they want to protect your IP. Just because you’re using AI in the platform doesn’t mean the information is getting shared to everybody else. Your IP is going to stay that way.”
Mike, a professional simulation expert himself, found the simulation AI session especially intriguing. He’s always required a lot of compute power to perform simulations, but with AI, “I can build these models that I can verify and do all this other stuff to, and then I can build an AI surrogate from that model and that AI surrogate can be used downstream by engineers and everybody else. So it’s a low cost approximation of that high cost simulation.” What’s most impressive is that the AI surrogates are getting more than 90% accuracy compared to the original models, and they’re running in minutes, not hours. The surrogates can also be feedback loops (or training) to the model for improvement. Mike says, “You have to be very careful about that because if you build in some sort of bias into the feedback loop, you could be making the model less and less accurate. So they’re very aware of this and they have these other tools that can check the bias and things like that.”
Learn More About COExperience
The full podcast contains more discussion and details on a variety of topics, including how the automotive industry is using feedback loops with simulated crash tests that are perfect use cases for training AI, how Virgin Galactic’s need to reuse spacecraft means PLM extends all the way to operations via an integration with DELMIA Apriso MES, how Dassault is helping COE revive the COE forum to facilitate conversations within the industry, why Mike encourages anyone and everyone to volunteer for COE, and more.
Be sure to check out Stay Sharp Episode 81: COExperience 2025 Recap and join us each week for a new podcast.
Follow the Razorleaf Podcast, Stay Sharp in Digital Engineering on:
