Case study: How Nestlé USA used AI to improve spare parts search and inventory visibility across its network of factories

At the 2025 Society of Maintenance and Reliability Professionals (SMRP) conference, leaders from Nestlé USA and SPARETECH outlined a maintenance transformation that began with a familiar frustration: finding spare parts in SAP.

“I’ll start with a typical story that I experienced pretty much every Saturday morning,” said Steven Gould, senior engineering maintenance manager at Nestlé USA. “As I’m sitting and drinking my coffee going through emails, I usually find an email or an IM or a text saying: Help. We need this part.”

Not all employees were familiar with how to search in SAP properly, and in some cases, it wasn’t possible for even experienced searchers to find what they were looking for. Corporate maintenance would need to log into SAP, try to decipher the information provided, and search across the company's many factories for inventory. 

Since a major transformation earlier in the year, that scenario has largely disappeared. “I have not had to have that Saturday morning bump into my schedule to go find a part,” Gould said. The parts’ inventory got a structured makeover, focused on data standardization and workflow simplification with the help of an AI-enabled tool integrated with SAP.

The problem: Factory-specific SAP data silos create spare parts duplication

The project began with a hard look at how Nestlé USA managed spare parts across its network of factories. With different practices across different plants and little visibility network wide, Andy Goldinger, senior expert maintenance engineer at Nestlé, described the core issue with the parts database: “It really became SAP material numbers specific to a plant, not the manufacturer’s part number.” 

Without the manufacturer’s part number, the parts are no longer searchable in SAP. The manual process to build a new material part number in the database took too many steps, and it wasn’t getting done properly. Across multiple facilities this had cascading effects. “That’s why we would have 20 factories stocking the same part with 20 different material numbers,” Goldinger said.

The challenges extended beyond duplication across sites. “As we started digging into it, it wasn’t just duplicates across different factories,” Gould explained. “It was duplicates within a factory, two or three of the same parts at different locations.”

During COVID, inventory levels had risen in response to low supply and long wait times from suppliers. “We started raising our stock levels up to compensate for that,” Gould said. “We’ve never started to bring those inventories down.” The result was inflated working capital tied up in spare parts with limited visibility into what was truly needed or already available.

Adding to the problem was limited cross-site transparency and limited employee access across sites. “They couldn’t see the inventory to a sister factory,” Gould added.
From a data standpoint, analysis was nearly impossible. “It was hard to really see usages on a specific component, because everybody had it different in SAP, and they had different SAP numbers,” Gould said. That fragmentation also limited leverage in supplier negotiations because they didn’t have a good handle on true stock numbers and company-wide usage.

Standardizing spare parts workflow with AI tool

In addition to introducing new technology, Nestlé also needed to address process consistency for entering new component data. “Every factory had their own way of doing it,” Gould said. “It might have been Microsoft Forms. It might have been a Power Query. It could have been an old paper sheet that somebody filled out.”

The goal became clear: “We standardized that across Nestlé USA to say we have one way,” Gould said. The guiding principle was simple: one part, one reference.

That meant defining approval workflows, clarifying who could create new materials, and documenting key processes as a standard. As Gould noted, “This was a big challenge for us because everybody had their own process.”

The company also tightened control over new material creation. Only designated approvers at each plant could finalize entries, usually the storeroom supervisor and the maintenance manager. “Only those two people can approve it,” Goldinger said. “You can’t push through stuff that would create duplications.”

To address the search inefficiencies and duplication, Nestlé implemented the AI-enabled tool SPARETECH on top of SAP. When trying to create a new material number, the AI system works by checking for potential duplicates as soon as a user begins entering a manufacturer part number. “The system is actually looking based off of what we already have in SAP for a match,” Gould explained. The duplicates show up at the bottom of the screen. 

The new system also provides more visibility into discontinued parts. If the factory doesn’t have an obsolescence plan or knowledge of what’s going to be replaced with the newest parts, SPARETECH can provide that support. It identifies discontinued components and flags future obsolescence dates. “It tells you what’s been discontinued, but also give you a future date if there’s something that is truly going to be discontinued in the future, so you have some time to react to that change.”

The tool also generates standardized descriptions automatically. “It is actually using AI to create the part description in SAP,” Gould said. Search functionality was another breakthrough with the AI tool. “It’s like a Google search,” Gould said. Rather than requiring complex wildcard formatting, technicians can type what they know and refine results dynamically.

Goldinger emphasized the practical value for technicians. “Your technicians don’t know SAP material numbers by heart, do they? That’s not what’s written on the gearbox out there on the information tag. That’s the manufacturer’s part number,” he said.

The tool also integrates catalogs with technical specifications and images. For maintenance teams working from tablets on the plant floor, visual confirmation is critical. “They can see a picture and actually match it up, before they make the trip back to the store,” Gould said.

Measurable results: Improved spare parts visibility and reduced inventory risk across the factory network

Based off some time testing, Nestlé saw major improvements in the measured time to find parts, about 50% faster, Gould said.

Beyond speed, collaboration has improved. “When it comes to that transparency of inventories, the collaboration has been huge for us when it comes to tech stores,” Gould said. Teams now communicate across sites to locate inventory and improve master data quality. “It’s not about what’s happening now in my store. I know my sister factory has it,” he added.

Each factory doesn’t need to stock every part at each facility. “We don’t need to stock those critical spares that are hundreds of thousands of dollars that are sitting on our self, just waiting to be used at some point, maybe, in the future,” Gould said. “We can now start looking at that inventory across markets.” Working capital reduction is now an active strategy for Nestlé. Rather than stocking high-cost critical spares everywhere, Nestlé can centralize inventory at selected facilities and transfer parts as needed.

“Having this visibility has allowed us to look at inventories as a market versus that individual factory,” Gould said.

Training and governance sustain new MRO data strategy

Nestlé approached the rollout as a structured change initiative, and the simplicity of the tool accelerated adoption, Goldinger said. “After about five minutes of playing with it, you really learn it,” he added. Super users were selected at pilot plants and then trained to lead subsequent waves. That peer-driven model helped sustain momentum.

Using a Microsoft Power BI, Nestle extracts the data to monitor compliance with the new workflow. “Are people using the tool? And we review that now in a lot of our network calls that we have on a monthly basis with the factories,” Gould said. On average, use of the tool exceeds 95% on a monthly basis. “They’re using this tool because they see the value,” Gould said.

The broader goal is to use standardized material numbers and clean data to support smarter stocking decisions across the network. Factories are no longer operating as isolated storerooms. They are part of a visible, collaborative network with standardized data and shared inventory insight. As Gould summarized, “This tool has really helped us make a lot of different people’s jobs in our factories much easier.”

For maintenance and reliability leaders, master data discipline is an important maintenance strategy, not just shoved off to IT. By standardizing spare parts data and improving visibility across sites, Nestlé is shifting from plant-by-plant inventory management to a coordinated network strategy that supports getting maintenance teams parts faster and better inventory management across the company.