Not long ago, I walked into a bank branch in the United States and was greeted by a robot. A real, physical robot, standing there welcoming customers as they arrived.

The scene stayed with me because it captured something remarkable about banking technology. Financial institutions have invested enormously in transforming the customer experience, from mobile banking and digital onboarding to real-time payments, fraud detection and, increasingly, AI.

And then you look behind the counter and find another technological era. Green screens are still being used to register transactions, legacy applications are connected through layers of interfaces, and back-office platforms that have been running for decades remain critical to everyday operations.

There is nothing irrational about this contrast. When technology investments compete for capital, the customer usually has a stronger business case than the General Ledger. Technology that can increase transactions, improve service, open new channels or generate revenue is easier to prioritize than replacing an ERP that continues to close the books every month.

So back-office modernization gets postponed. There is always something more urgent. The answer becomes “not now.”

The problem is what happens after years of “not now.” The business keeps changing, new applications require integrations, acquisitions introduce additional systems, reporting requirements generate workarounds, and manual reconciliations compensate for gaps between platforms. Each decision may make perfect sense on its own, while collectively they create an increasingly complex architecture around technology that becomes progressively harder to replace.

The more interesting question, then, is not whether the old ERP still works. It is what the organization is giving up because it still works that way.

An ERP Implementation Done Backwards

I saw one Financial Services company answer that question in an unusual way.

Its U.S. headquarters was still operating on an older ERP, while one of its Latin American subsidiaries had grown dramatically during the pandemic. The business had multiplied its channels and transaction volumes, and the local Finance operation needed to support a scale for which its existing back office had not been designed.

The conventional decision would have been to extend the headquarters ERP to the subsidiary and leave modernization for another day.

They did the opposite. Instead of bringing the old headquarters back office into a fast-growing business, they implemented Oracle Fusion Cloud ERP in the Latin American operation first. In other words, they modernized from the outside in.

The results went well beyond having a newer ERP. Finance was able to adopt more standardized processes instead of reproducing years of historical customization. The company gained stronger integration capabilities, particularly important in Financial Services, where ERP must coexist with multiple transactional and specialized applications. It gained better analytics and a platform that could continue incorporating automation and AI capabilities as they evolved.

Most importantly, the back office became capable of scaling with the business rather than forcing the business to keep building around its limitations.

When we think of back-office modernization in this industry, we typically compare the cost and risk of replacing an existing ERP with the cost of keeping it running. As long as the system remains stable, postponement can continue to look like the financially responsible decision.

But keeping a system running and keeping a business moving are not the same thing.

The real cost may be found elsewhere: in the difficulty of integrating a new business or channel, in Finance teams spending valuable time on reconciliations and workarounds, in the additional administrative capacity required as transaction volumes increase, in fragmented financial data, or in the effort required to introduce automation and AI across processes built on decades of applications and integrations.

For years, putting customer-facing technology first made considerable business sense for financial institutions. In many cases, it still does. The question is whether back-office modernization must continue to wait until replacing the headquarters ERP becomes an enterprise-wide priority.

This experience suggests another path. Perhaps modernization does not always need to begin at headquarters.

Starting with a smaller country, subsidiary or business unit can make the scope and investment more manageable while allowing the organization to begin building its future back-office architecture. One location can be modernized, then another, gradually moving the organization forward without requiring a massive transformation from day one.

In that sense, this implementation may have been done backwards, but perhaps that was precisely the point. Headquarters does not necessarily have to be the first step. It may eventually become the last.

Perhaps the question for CIOs is no longer whether back-office modernization should take priority over other technology investments, but whether modernization really needs to start at the center, and what the business may be giving up by continuing to postpone it.