Business
SAP : AI Agents to Take Over 100,000 Manual Order Confirmations at Lemvigh‑Müller
SAP : AI Agents to Take Over 100,000 Manual Order Confirmations at
About this update from Sap Se
The Danish wholesaler Lemvigh-Müller has deployed artificial intelligence to automate one of the most time-consuming tasks in procurement: processing supplier order confirmations. The solution consists of multiple AI agents, each responsible for a clearly defined task, orchestrated into a single automated workflow built on SAP Business AI. The outcomes are faster processing, improved data quality, and more accurate delivery information for customers. When suppliers send order confirmations as PDF files, even minor discrepancies in price, quantity, or delivery dates can trigger significant manual effort within procurement. For Lemvigh-Müller, one of Denmark's largest wholesalers within steel, plumbing, heating and electrical products, this has long been a familiar challenge, consuming substantial time and resources. The company has now tackled the very point where earlier automation initiatives often stalled. With a new solution based on several specialized AI agents, developed on SAP technology and implemented in close collaboration with NTT DATA Business Solutions, supplier PDF order confirmations can now be read, interpreted, compared, and processed automatically-directly against SAP systems. Capture business-wide AI value with speed and confidence Learn more "We have previously tried both RPA and traditional automation approaches without really achieving the desired effect. The key difference this time is that we broke the task down into multiple independent AI agents, each responsible for a specific part of the process. Together, they now handle what previously required manual review," says Frederik Aakerlund, IT director at Lemvigh-Müller. 10 weeks from idea to AI agents in production The project originated with an e-mail from Jess Frederiksen, an AI-savvy project manager in Lemvigh-Müller's Market and Procurement organization. After successfully matching an order confirmation with a purchase order using ChatGPT as an experiment, he approached the IT director to explore whether this could be turned into a fully integrated system solution. From the initial tests to production deployment, the entire project took just 10 weeks. According to Lemvigh-Müller, this short implementation timeline was critical in allowing the solution to demonstrate tangible business value quickly and build internal support. "This was not a long-running project. In 10 weeks, w...