Where ERP transformations struggle and how AI can help

By Peter Wienand, Elisabeth Goos and Ralph Seidel

Large-scale process and enterprise resource planning (ERP) transformations are often viewed as expensive, carrying high risks of failure. Indeed, companies undertaking these large-scale projects face challenges from multiple complexities.

From a functional perspective, diverse processes, silos, and varying maturity levels can complicate the start. From an organizational standpoint, complex structures, resource and skills gaps, inefficient decision-making, and cultural blockers add to the complexity. Lastly, there are usually massive technical challenges, ranging from fragmented architecture and decentralized IT systems to existing technical debts.

AI can help. It can serve as a catalyst for transformation programs, making them faster, more robust, and cost-effective. And AI can be directly embedded into end-to-end processes, maximizing the transformation’s impact on value.

A supercharger for transformation programs

AI offers support throughout the entire transformation lifecycle. It establishes clear guardrails by analyzing vast amounts of data to shape a coherent vision and mission while considering the business' unique characteristics. Then it can accelerate key deliverables, such as user stories, process specifications, and documentation, while facilitating smoother handovers between business and development teams. AI also enables faster development through automated code creation, debugging, and enhanced technical specifications. Finally, in testing, AI speeds up the generation of test cases and documentation as well as increasing the number of test scenarios to improve system robustness, while avoiding repetitive tasks for humans.

Given the growing economic pressures on companies and ongoing transformations, it's clear that AI has become a versatile tool for reducing transformation costs and complexity. While the benefits are clear, many specific AI tools are still in prototype stage and require further maturing to become a standard component of transformation programs.

Embedding intelligence in core processes

While accelerating project work is one relevant value driver, embedding AI directly into the desired processes helps address the last bastions of repetitive manual work that were difficult to tackle with conventional automation tools. AI therefore offers a sustainable way to impact daily operations and create lasting value. This integration occurs when shaping and redesigning future processes, from standardized and automated processes like source-to-pay, to more complex ones such as plan-to-produce and innovate-to-market.

For example, in accounts payable as part of source-to-pay, AI agents can instantly match supplier invoices with purchase orders and goods receipts, flagging price mismatches, and routing approval processes. In spending control, AI can identify if teams in different countries are buying the same items from various suppliers at different prices and detect spending patterns for optimization through consolidation or timing adjustments.

Standardization and harmonization in production present different challenges, but AI also moves beyond simple automation here. An inventory AI agent can analyze past sales, seasonality, and supply lead times to determine optimal safety stocks per item, thus optimizing working capital. AI can also be used for predictive maintenance, rerouting production to avoid costly downtimes. Furthermore, AI-supported forecasting processes can incorporate various data from sales, marketing, and distribution partners to enhance demand forecast accuracy.

In innovate-to-market, AI can boost simulation capabilities to predict performances of several product prototypes or variants, drastically reducing the need for costly and time-consuming physical testing. It could also help with the discovery of new, less resource-intense material combinations based on the latest research while considering internal specifications and quality standards.

Integrating AI into future end-to-end processes is a critical component of transformation, especially as processes may change depending on AI's capabilities. Anticipating further enhancements post-transformation is essential, leveraging AI agents to analyze process data for continuous improvement. ERP providers have already taken steps to include production-ready AI functionality in their product suites. To maximize AI's potential, they must facilitate workflows that combine human and AI input and facilitate contextualized data as a baseline for AI agents. Decisions taken in ERP transformations will impact a large share of company data and with that the foundation for agentic AI, including its cost and reliability.

Accelerate transformation and increase value

Large-scale transformations are inherently complex and require substantial investment. AI offers a means to:

• Enhance and accelerate transformations by selectively supporting key tasks and minimizing manual work
• Significantly boost value generation in future process design when integrated early in the design phase

At Roland Berger, AI now plays a valuable role in our transformation program portfolio. We use it to rapidly define processes, for instance, or to generate value cases that are industry- and business-model-specific.

Organizations can benefit twice by enhancing their transformation efficiency through targeted use of AI within weeks, as well as by shaping a best-in-class process landscape with potential to continuously learn and adapt.

Sign up now to receive regular insights on digital transformation and value creation via e-mail.