I have sat through enough tool pitches to manufacturing engineers to know the usual fate: a polite nod, then quiet neglect. Engineers are busy and skeptical, so they guard how they work for good reason. When I tell you LEO is worth your attention, I want to make the case plainly. This is with the numbers we are seeing and without the launch-day noise.
LEO is one of the AI virtual companions Dassault Systèmes has built into the 3DEXPERIENCE platform. The distinction that matters is that it is not a chat box bolted onto your work. It reads the design, the process context and the data already in the platform. It acts on that context instead of just describing it. It works across engineering, manufacturing and simulation, so the same companion that helps reason about a part can help define how that part is built. It will then validate the result virtually before anyone cuts metal. That span is where the most expensive disconnects in a factory have always been hidden.
Why Are Manufacturers Turning to AI to Mitigate the Loss of Tribal Knowledge?
Most of your manufacturing know-how is not written down. It lives with a handful of senior engineers and operators. It is with the person every new hire shadows and who knows why line 3 needs a fixture tweak. That knowledge is leaving. Close to a quarter of the U.S. manufacturing workforce is 55 or older and when an experienced engineer retires, as much as 70 percent of their undocumented knowledge can walk out with them. Replacing it is slow: six to nine months for a new hire to reach full productivity and up to eighteen months in roles built on pattern recognition.
Defining a process is still slow and manual. An engineer reads the design, sets the operations and sequence, picks resources, writes instructions and checks against plant rules that may not be documented. Most of it is re-derived from scratch, even when a near-identical process existed last year. The result is long lead times, uneven quality and an organization one retirement away from forgetting how it makes something.
What Are the Core Efficiency and ROI Metrics of the LEO Virtual Companion?
Strip away the marketing and LEO does three things. It reveals the manufacturing context, constraints and dependencies an engineer would otherwise hunt for. It performs the repetitive MBOM, process and work-instruction tasks that used to eat an afternoon. And it generates new MBOMs, processes and instructions from the design and captured know-how.
The numbers we put in front of operations leaders are specific:
- +50% manufacturing engineering efficiency
- -25% industrialization lead time
- -50% cost of non-quality
- ≈1 FTE added capacity per engineer
A task sequence that used to take two and a half hours—searching documents, recreating the MBOM, interpreting standards, and updating instructions by hand—now runs in under ten minutes because the engineer asks rather than rebuilds.
This is not automation in the replace-a-person sense. It is closer to giving every engineer a fast, well-read junior who has absorbed the whole plant’s process history and forgets nothing. The senior engineer still decides. LEO removes the hours of searching and retyping between the decision and the result.
How Does LEO Automate the Complex EBOM-to-MBOM Transformation Process?
If you want to see where a companion earns its place, look at the manufacturing bill of materials. The engineering BOM says how a product is designed. The manufacturing BOM says how it is actually built, in sequence, with the resources and steps the design never mentions. Getting from one to the other is a transformation, not a copy and it is hard.
Across 173 discrete manufacturers, that move reorganized about 76 percent of components and added roughly 195 manufacturing-specific items that did not exist in the design. It consumes 12 to 18 percent of new-product-introduction resources on average, nearly 24 percent in automotive and 26 percent in aerospace. And it does not sit still: the MBOM absorbs more than eight revisions in the first year of production alone.
LEO works the MBOM in three moves. It generates a draft MBOM from the design and captured knowledge, structured the way your plant builds, so days of restructuring become a draft you refine. It reconciles by continuously checking the as-designed state against the as-built state and flagging the mismatches, duplicates and unit conflicts behind nearly three-quarters of quality issues, turning a periodic fire drill into a background process. And it manages change, tracing an engineering change into the affected assemblies, steps and instructions, applying it with the right effectivity and preserving traceability. None of this removes the engineer. It removes the part of the MBOM job that was never really engineering.
Why Is AI-Guided Process Engineering Critical for EV and Aerospace Manufacturing?
The case is strongest where products are complex, variants are many, and errors are costly. EV makers already spend close to a quarter of NPI effort on the EBOM-to-MBOM transformation and EVs make it worse: the bill of materials is being reinvented around batteries and software-defined systems, variant counts are climbing and cycles have compressed beyond what manual definition can sustain. When a battery module is redesigned mid-cycle, the change has to reach the line in days. We are not theorizing here; Hyundai has worked with us on the platform to develop optimized EV battery packs and lightweight frames.
Aerospace and defense (A&D) sits at the other end: low volume, extreme complexity, deep configurations and a heavy traceability burden because certifications and lives depend on it. Aerospace carries the highest transformation cost of any sector, around 26 percent of NPI resources and in one survey, 41 percent of A&D organizations were already piloting generative AI in 3D modeling. Here, the value is less about raw speed and more about consistency and proof: capturing configuration knowledge that lives with a few veterans, tracing change with auditable precision and validating assembly sequences in simulation before a fastener is placed. A process proven in the virtual world is one you can defend to a customer and an auditor.
What Data Requirements and Infrastructure Are Needed to Implement Manufacturing AI?
I would not serve you well by pretending this is effortless. A companion is only as good as the data and captured knowledge beneath it. If your process history is a sprawl of disconnected spreadsheets, LEO will help, but the first real work is getting that foundation into the platform. The productivity figures are genuine but context-dependent; treat them as a direction, not a promise. And the engineer stays in charge. The value is not removing human judgment; it is removing everything around that judgment that wastes your best people’s time.
How to Scale Generative Process Planning with the DELMIA Value Engagement Discovery Offer
Everything above is LEO assisting task by task. That is valuable now, but it is the on-ramp, not the destination. The destination is the generative process plan: you describe the intent and the plant while the system drafts the whole process, the sequence, the resources, the MBOM, and the instructions. The engineer moves from author to editor.
If you want to see this on your own products rather than in a demo, that is what the DELMIA Value Engagement Discovery Offer exists for: a focused approach that runs LEO Process Engineering inside one of your own virtual twins, so you see the value on your own products rather than in a demo. If that is interesting, approach us and we will work out the right path together, the use cases worth starting with, the data to put in place and how we measure the value as we go.
Every quarter you wait, more of your plant’s hard-won knowledge walks out the door and it does not come back.
DELMIA, a Dassault Systèmes brand, connects the virtual and real worlds to drive innovation and sustainability. Powered by the 3DEXPERIENCE platform, our end-to-end solutions integrate virtual twins, industrial AI and augmented reality to optimize manufacturing, supply chains and workforces. We empower businesses to reduce waste and achieve sustainable, customer-focused operations, building a more resilient future.
Prashanth MysoreSenior Director for Strategic Business Development, DELMIA