Industry: Metal (Steel) and Foundry
Category: Artificial intelligence, SAP S/4HANA

From repetitive work to real impact with AI

This is the first GenAI case we are building with full SAP S/4HANA integration, and we already have more on the drawing board. It has definitely given us an appetite for launching new projects.

Frederik Aakerlund CIO, Lemvigh-Müller

About the customer

Lemvigh-Müller is one of Denmark’s leading B2B distributors of steel and technical installation products. The company manages thousands of suppliers and complex procurement processes across its supply chain.

Learn more: https://www.lemu.dk/en

Business needs:

  • Reduce time spent on repetitive, manual validation of supplier confirmations
  • Handle large volumes of unstructured data in a scalable way
  • Improve efficiency without replacing employees, freeing them for higher value tasks

Solution – An AI-driven solution fully integrated into SAP that:

  • Extracts and structures data from supplier confirmations and matches them against purchase orders
  • Identifies discrepancies and triggers automated workflows or email responses
  • Provides full visibility through a monitoring dashboard
  • Is built using SAP-native technology and the latest AI framework, following SAP’s new delivery model

Outcome:

  • Touchless processing rate reached ~93 % within just a few weeks of going live, surpassing the initial end target of ~80–85 %
  • Significant reduction in manual workload
  • Faster and more accurate handling of supplier data
  • ~23% of suppliers onboarded within the first few weeks
  • ~98% average SAP Match confidence score, reflecting high agent accuracy

Why NTT DATA?

  • Strong experience with SAP Business AI, SAP BTP, and complex SAP backend integration, making it possible to design AI solutions that are tightly aligned with the existing SAP landscape.
  • Supports clients move quickly from proof of concept to scalable production.
  • Ensures a setup that is governable, secure, scalable, and aligned with enterprise compliance requirements.
  • Works closely with business and IT stakeholders to ensure AI solutions is not only technologically sound but also based on real business needs.

Business need

Thousands of unstructured confirmations created operational bottlenecks

Lemvigh-Müller receives thousands of supplier order confirmations each month from more than 35,000 suppliers, often in unstructured formats such as PDFs and emails.
Each confirmation had to be manually reviewed and matched against SAP purchase orders: a time-consuming and repetitive process with limited business value that slowed down operations.

The goal was not to replace employees, but to remove repetitive workload and enable them to focus on exceptions, supplier dialogue, and decision-making. Using AI as a tool, not a replacement and freeing employees from low-value work to focus on what matters.

Solution

AI handles the routine, people handle the exceptions

NTT DATA implemented an AI-based solution that supports employees by activating existing business processes quickly and efficiently. Users remain fully in control through a dashboard, where they can review, adjust, and approve decisions as needed.

The solution is built on SAP’s latest AI framework and delivery model, fully integrated into the existing system landscape without the need for external tools or complex setup. Acting as a digital assistant, it processes large volumes of repetitive tasks while maintaining compliance and governance.

From proof of concept to live operations in just weeks, the solution automatically reads and structures supplier confirmations, matches them against purchase orders, and flags only cases requiring human attention. Rather than replacing employees, it removes repetitive work and
enhances decision-making.

Developed entirely within the SAP ecosystem using the latest technologies, the solution is fast to implement, easy to scale, and ready to support future AI use cases.

Through close collaboration and rapid iteration, it was delivered as a lightweight, scalable solution in 6–8 weeks with minimal setup and no need for large consulting teams—demonstrating how quickly business value can be realized. With this foundation in place, similar solutions can be deployed even faster.

Outcomes

From manual effort to automated flow

The solution transforms a previously manual, time consuming process into a streamlined, automated flow where the majority of order confirmations are handled without human intervention.

Instead of reviewing every document, employees are only involved when deviations occur, allowing them to focus on resolving exceptions and adding value where it matters most.

This shift not only improves efficiency but also ensures better scalability and a more resilient supply chain operation.

At the same time, the solution demonstrates how AI can quickly create tangible business value — not as an experiment, but as a practical tool embedded in daily operations.

All of this happened in very close collaboration, and that has been one of the really strong parts of the project. In the project the developers had direct access to the business and our SAP resources, so we were able to create solutions quickly.

Klaus Heinemann Head of SAP ERP, Lemvigh-Müller

Read more