Manufacturing Voices: Ben Snyder, Vice President, Lifecycle Services, nVenia

nVenia builds highly engineered-to-order equipment across five brands, where closely similar parts had proliferated across internal brands faster than manual searches could reconcile them. Ben Snyder, Vice President of Lifecycle Services, evaluated engineering AI early, paused the investment when the technology was not yet ready, and re-engaged CADDi once it had matured. In this VOICES article, Design World spoke with Snyder about controlling SKU proliferation, reducing inventory risk, and proving out engineering AI in a production environment.

Design World: Tell us a bit about yourself and your role at nVenia.

Ben Snyder: I am Vice President of Lifecycle Services, which is our aftermarket department. I’m relatively new in that role, having started in February of this year. Prior to that, I was in the operations role at nVenia for two and a half years. Thanks to my time with operations, I already had a good understanding of the execution and operational excellence side of the business, and now I also understand this from the commercial side dealing with questions like how we’re using new tools, such as CADDi, to really drive revenue increases at the same time as we streamline efficiencies inside our factory.

nVenia builds highly engineered-to-order equipment across five distinct brands. How does the engineering team fulfill the core business mission for your customers when managing complex product variations across internal brands?

We build a combination of what we call engineered-to-order (ETO) style machines, as well as configured-to-order machines.

Configured-to-order is easier to describe. It’s like ordering a car. You have a set number of options, you select the combination you want, and then we will go and build your unit. The difference is that we have 30,000 to 45,000 different options, so there is a lot to choose from.

On the ETO side, these are unique opportunities that come in. For example, a ramen company or a bottling company. We have machines that have a common core, but every product is slightly different in size. This requires significant customization, for how that specific product comes to the machine, how it gets bundled and packed, how we wrap the product and place it onto a pallet with a robotic palletizer, how that gets sent to a semi-truck or storage area.

All of these unique customer opportunities require special components. Prior to CADDi, the idea was that we must build a new part, make a new SKU, release it to the factory and go from there.

What we learned doing manual searches is that we have many components that are very similar. A part can have the same shape but the hole is in a different location, or we might find two parts with an identical shape but one is not used because someone in the engineering department designed it 15 years ago and newer engineers don’t know about it. We use CADDi to help us locate those similar parts so that we don’t have SKU proliferation.

Engineers spend roughly an hour a day searching for fragmented parts data. What specific operational bottleneck made eliminating SKU proliferation the primary catalyst for evaluating a new software solution?

That goes back to what I said earlier, and the idea that every time you produce a new part, there’s a lot of work that goes into that. It’s not just creating a drawing. It’s needing stock and a stocking strategy. You have to set it up in your enterprise resource planning (ERP) system. There are quality inspection steps. If we’re buying this from outside, I need a procurement plan. I need a ton of different things that come into play just to create a nut, for example.

We wanted to stop that confusion. We have hundreds of thousands of drawings inside of our facility. How do we optimize using them over anything else before we create something new?

nVenia currently holds sizable inventory to allow for decreased lead time with the many combinations of product. How will centralized part data help reduce excess working capital while generating new project revenue through component association?

The first thing we wanted to do is stop the bleeding. In order to do that, don’t make new parts, to avoid SKU proliferation. We’re using CADDi on the design side, and we’re only producing new parts when we absolutely must.

The second thing is dealing with what we call excess and obsolete (E&O) inventory. What we want to do is be able to utilize those parts in a new design. When we get these ETO machines coming down the pipe, we want to design in those existing parts. Instead of using a 1.2 horsepower motor, we can use 1.3 horsepower because that’s in inventory, so we can save some money there.

The last thing is what happens when a customer calls. We have this inventory because it’s been used somewhere; we’re not just buying parts that never get used. Typically, a customer has a custom machine and the part might only break every five or 10 or 15 years. So it becomes about us reminding the customer. What CADDi helps us do now is call customers to say, this is an at-risk inventory part. We can give you a discount if you buy two of these, or 10 of these. It flags us as a customer service representative, to make sure we’re letting customers know that we don’t plan to stock this and there’s a great opportunity to put some safety stock of this part on their shelf.

So we’re stopping the bleeding, and not producing more SKUs. We’re using CADDi internally, so I have a list of parts that engineers can search where they can say, I need something that looks like this or I need a motor that does this. CADDi will suggest ways to use up existing internal inventory.

Finally, we want to sell it. When a customer calls in for a spare part order, we’re able to identify the part quickly and say, ‘We have that and we have more in stock. Would you like some?’

You previously paused engineering AI software investment because the early technology lacked a direct link to your engineering vault. What specific real-time synchronization capabilities convinced your team to reconsider the CADDi platform as a mature system ready for nVenia’s needs today?

We engaged in CADDi about two years ago, and did a year-long trial. During that trial, we figured out that due to the engineered-to-order nature of our business, we were syncing about once every week, maybe once every two weeks, and that was too slow for us. We did this as a manual sync. Data would come in, but in real time when we’re designing something I need to find something new now, not two weeks from now.

The game changer came when CADDi developed an API that allowed us to transfer this data almost in real time. That also allowed us to use another tool called CADDi Quote. Before, if drawings would come on Monday at noon, and I needed to quote them by three or four that afternoon, I couldn’t. Now we can.

The core of CADDi is that I put in all the drawings and then I can look at those drawings through the AI filter. I can ask, ‘find me a pipe with this diameter’ and it’ll go search, or shape match, and find me something that looks the same. On top of that is the CADDi Quote add-on module. When we look at going out to our vendors and saying, ‘I need to manufacture this outside of our four walls,’ it’s to ensure we’re getting good pricing. We wanted to use this tool to help our procurement team get the best possible cost.

These days, we get those new parts in real time. The procurement team gets a notification to go bid these out. We bid them out, usually to three manufacturers, and see which ones get the best price. We can also pull in quality data from that manufacturing group. We can say, ‘I got a good price from a good quality manufacturer—let’s move forward, let’s issue a PO to them.’ CADDi Quote really helped us on that front more than anything.

Evaluating a returning software vendor requires clear evidence that their system can handle high-volume production data. What specific results from uploading 3,000 parts at once confirmed the software had evolved enough to process your real drawings accurately?

They have approximately 750,000 drawings of ours right now. 3,000 were from doing the initial demo, which was really interesting.

We grabbed 3,000 random parts, gave them the files, and they were able to find some neat information about commonalities. The biggest one was a right-angle bracket that we use to hold up sensors. Randomly in those 3,000 parts, we had three or four matches. The only difference was that the holes were in different locations.

So the idea became, can we put all four holes in this at the same time so I only have one part? Or is this a modification where I can take the three parts down to two, or can I slot it instead? It raises a lot of questions about why we have so many similar parts. Just in that quick demo of 3,000 drawings, we found amazing opportunities.

CADDi is a global technology company that develops a vertical data intelligence platform for engineering, design and manufacturing. CADDi brings AI technology expertise to help manufacturers preserve decades of engineering knowledge. Rather than replacing PLM or ERP systems, CADDi operates as a shared intelligence layer across manufacturing teams.

CADDi’s new white paper explains why improving decision velocity can drive significant productivity gains without large capital investments, while also improving the quality of decisions. CADDi will continue expanding its platform to help manufacturers strengthen operational performance, reduce friction in engineering and sourcing processes, and build the intelligence foundation needed for long-term competitiveness. More case studies and insights are available at us.caddi.com/case-studies.