The Hidden Cost of ERP Customization

For years, IT leaders have extended or modified enterprise resource planning (ERP) systems to match unique business processes. This approach worked—until now.

As artificial intelligence (AI) becomes increasingly integrated into operations, a clean, standardized ERP core is emerging as the foundation for success. Angela Maragkopoulou, CIO at Sunlight Group Energy Storage Systems, frames this need clearly: “Don’t sell standardization to your board as technical cleanup; frame it as the price of admission to AI.”

The Growing Gap Between IT Leaders

While AI investment accelerates across industries, a significant minority (51%) still favors customized ERP approaches—a figure that hasn’t changed much in recent months. This creates a competitive gap between those addressing decades-old technical debt and those who haven’t yet prioritized the issue.

James Baker of Bain & Co. notes that companies aren’t yet “funding wholesale de-customization purely for AI.” But this decision carries significant implications, as AI grafted onto complex systems inherits every flaw in the foundation.

Why Customization Happened—and Why It Matters Now

Customization initially addressed legitimate needs when ERP systems didn’t offer native functionality. As Maribel Lopez of Lopez Research explains, “ERP companies have been trying to get customers to move to the cloud for years, but customization created a lock-in.”

But what changed is who—or what—is using these systems. AI agents require reliable data and standardized processes; customized environments create fragmented versions of the truth that undermine trust.

Think of it like this: customizing an ERP system is like building a bespoke bicycle. While functional, it’s less efficient for others to use and harder to upgrade or maintain.

The First Signs of Impact

The challenges are most apparent in high-volume, rules-based processes where AI offers clear benefits—finance close, procurement contract management, and supply chain demand planning.

Sunlight Group’s experience illustrates this point. Maragkopoulou standardized their S/4HANA implementation primarily to address end-of-support issues and improve reporting—with AI becoming an unexpected benefit later on.

“A clean foundation makes AI possible, not automatic,” she emphasizes. By cutting unnecessary customization and focusing only on genuine differentiators, Sunlight created a system where data is consistent and trustworthy.

A Clear Choice for Future-Ready IT

Maragkopoulou’s advice to CIOs is straightforward: “Standardize first, or don’t bother.” Because when you automate complexity instead of solving it, you amplify your problems rather than eliminate them.