Output or prerequisite?

What specific AI use cases in business should companies actually be pursuing today? The short answer is that many of the most valuable AI use cases can and should be introduced before a large-scale ERP transformation - not after. This article is an attempt to explain why.

Using AI - but how?

The instruction to „use AI“ has become almost ubiquitous. Most of us already rely on it for day-to-day administrative tasks, and at Lemongrass we use AI in operations, support and even project delivery. But we are a technology-centric company. Many large companies operate with varying degrees of complexity, legacy systems and organizational inertia.

Due to our roots, we naturally focus on SAP-related AI use cases. The long-term vision, where AI is ubiquitously embedded in every level of a business process, silently optimizing, predicting and automating, is extremely compelling. Making supply chains more robust, improving demand forecasting, maximizing production output and increasing sales with minimal increase in distribution costs presents huge opportunities. So it's no surprise that every boardroom is talking about it.

What many organizations fail to do, however, is to assume that achieving this future state will first require a comprehensive ERP transformation. Yes, clean, modern and well-integrated ERP landscapes facilitate the use of AI. No one disputes that. But significant, high-value AI opportunities already exist today - without a multi-year S/4 migration, a greenfield reboot or a complete master data overhaul. In fact, many organizations already have enough basic „infrastructure“ in place to realize tangible, AI-driven value. These initiatives can deliver value quickly and, importantly, help to fund and underpin the more complex transformations that follow. Let's take exception management as a simple example. Many companies employ large teams to handle invoice errors, purchase order discrepancies, policy violations, missing approvals and incorrect subtotals. This is an ideal use case for GenAI - and one that can be implemented much faster than most expect.

GenAI agent in use

One of our clients followed a similar philosophy by combining multiple sources of customer demand data with real-time visibility of inventory levels. By using a GenAI-powered sales agent to identify missing products and services on invoices, they were able to generate hundreds of millions in additional revenue. Again, fully integrated and with automatic updating of orders directly in their cloud ERP systems. These are pragmatic, high-impact use cases that can be implemented quickly. Nevertheless, many organizations remain convinced that they must first complete extensive, capital-intensive ERP programs - replacing legacy systems, cleansing all master data or replacing entire application landscapes - before they can begin to realize AI-driven added value.

Transformation first?

It is precisely here that the methodology itself deserves critical consideration: a „transformation first“ mindset treats AI as a downstream reward - as something that can only be earned after years of upheaval, costs and risks. An „AI first“ mindset, on the other hand, sees AI as a diagnostic and value creation tool.

While an „AI first“ approach is not risk-free, the pace of innovation and AI's rapidly growing ability to create, retain and reason about contextual information should at least give organizations pause before embarking on multi-year transformation programs. When will AI be able to capture and traverse networked business processes end-to-end? Probably sooner than many expect. Not because companies are perfectly aligned, but because AI is increasingly deriving these connections from observed behavior - rather than from prescribed design. Of course, this doesn't solve the complex issues of governance and company-wide alignment. But if the scope of the big „T“ transformation is reduced, these issues will also become more manageable.

In a perfect world, transformation would perhaps always come first. But we don't live in a perfect world. We live in a world that is changing and accelerating faster than ever before.

So the real question is: if AI can be deployed quickly across ERP systems and the entire ecosystem to solve key business problems, create measurable value and make visible where transformation is really needed - isn't that where our methods should start? (Source: Lemongrass)