Key Takeaways
- Master data automation works best when ownership shifts from IT to the business teams closest to the data, so master data experts can focus on governance instead of data entry.
- Workflow visibility, validation at the point of entry, and touchless posting to SAP are what move teams from desktop automation to full process orchestration.
- AI results depend on data readiness, and the strongest accelerator for any automation program is a champion inside the business.
Who should own your master data: IT, or the business teams that use it every day?
That question ran through the customer panel I moderated at the Automate Community on Tour event in Chicago. I was joined by:
- Jonathan Barrow, Senior Manager of Master Data and Product Information Management at Fender Musical Instruments
- Lindsay O’Hair, Director of Data and Analytics at CROPP Cooperative (Organic Valley)
- Srini Boppana, Director and Enterprise Architect Lead at Mondelēz International
Their companies differ in size, industry, and SAP maturity, yet their views on master data automation lined up more than I expected. Here’s what stuck with me.
Business ownership is a real driver of master data automation
Each panelist started in a different place.
Organic Valley went live on SAP S/4HANA in April 2024 as a Greenfield implementation, which meant rebuilding every master data intake process, many of them manual SharePoint-based forms. Lindsay told the room that Automate Studio first helped relieve that pain. The longer-term goal is to push ownership of master data to the teams that use it most, and to spend less time on data entry and more on governance.
Jon agreed, and he was blunt about what that means for his team. “They’re not data entry clerks, they’re analysts,” he said. Their job is to set rules and make sure those rules are followed, not to chase down approvals or double-check spreadsheets.
Srini described the same pattern from the Mondelēz side. His team began with finance master data maintenance, then expanded quickly as exception scenarios were built out during implementation. Other teams noticed and adopted the tool.
That fits how I see it at Precisely: the people closest to the data should own it.
Full workflow visibility helps you work faster, reduce errors, and track every step
With enhanced visibility, teams move from isolated desktop automation to orchestrated, end-to-end processes.
Fender has used Automate Studio for about 14 years. Jon’s team installed Automate Evolve with workflow as the draw. Fender’s material build process is complex, and seeing it laid out in swim lanes made it easier to manage. The team is starting with small workflows, pulling existing Automate Studio scripts into Evolve and adding workflow on top, before tackling new product introduction.
Lindsay named workflow as the reason for Organic Valley too. Her business users are still learning SAP’s terms and definitions, so she wants the process itself to teach them. Her goal is to “wrap governance right into the process of collecting that data,” so requesters understand the quality of what they submit.
Srini added a different angle. Mondelēz wanted touchless processing: once a workflow completes, the posting to SAP triggers automatically. Audit traceability was another reason.
Three customers, three approaches to AI
Lindsay shared that through conversations at a recent Gartner event, she picked up on three different approached to AI: go all in, wait and see, or try and see.
Organic Valley leans toward try and see, with a cautious eye toward protecting its farmers. Her main lesson from the proofs of concept: “If the data is not ready, then it doesn’t matter which tools you use or which processes you use, it’s not going to work the way people intend.” Rather than hunting for use cases, her team lets them surface through existing requests, then prepares the priority data behind them.
That echoes what I said on stage: great automation on bad data just automates bad data, and AI amplifies the problem at scale.
Jon is further along. His first AI project built a vector database from roughly 12 years of product copy stored in the product information management (PIM) system, so product managers can draft copy in each brand’s voice. He has since built an AI agent called Leo, connected to Fender’s EnterWorks data, that generates product recommendations. Jon said Leo has produced 119,000 recommendations since December, compared with 55,000 that a team of five created over three years. His takeaway from a flawed demo that became a new rule within seconds: “You’re inspired by the failures.”
Srini shared that Mondelēz already has AI use cases in place, such as knowledge assistance, and is collecting and prioritizing ideas that deliver quick benefits. The foundational data layer is a key part of that thinking.
What would they tell teams just getting started with master data automation?
I closed by asking what each panelist wished they had known at the beginning of their automation journey. The answer was nearly unanimous: find your champion.
Jon described the ideal advocate as someone detail-oriented who enjoys making order out of chaos. Lindsay said the most movement happens when someone in the organization is excited about how the work improves their day, and she noted that AI is already shrinking some development work from a week to an hour. Srini agreed on the value of a champion. Once a team understands the tool’s capabilities, he said, dependency on IT drops, and adoption rises.
Lindsay’s S/4HANA go-live lesson: think about master data processes earlier. Srini noted that Mondelēz is migrating from ECC to S/4HANA and looking ahead to Fiori-based transaction processes.
Master data automation success starts with people
The most consistent theme from Chicago wasn’t a feature. It was people: business owners who take responsibility for their data, champions who build momentum, and master data teams recognized as analysts. The technology works best when those pieces are in place first.
See how Automate Evolve can bring workflow and governance into your SAP master data processes.
