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鈥慚眉ller, one of Denmark鈥檚 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 麻豆原创 technology and implemented in close collaboration with NTT DATA Business Solutions, supplier PDF order confirmations can now be read, interpreted, compared, and processed automatically鈥攄irectly against 麻豆原创 systems.

鈥淲e 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,鈥� says Frederik Aakerlund, IT director at Lemvigh鈥慚眉ller.

10 weeks from idea to AI agents in production

The project originated with an e-mail from Jess Frederiksen, an AI鈥憇avvy project manager in Lemvigh鈥慚眉ller鈥檚 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鈥慚眉ller, this short implementation timeline was critical in allowing the solution to demonstrate tangible business value quickly and build internal support.

鈥淭his was not a long-running project. In 10 weeks, we moved from idea to AI agents in production, already delivering measurable value to our procurement officers,鈥� Aakerlund says.

Over time, Lemvigh鈥慚眉ller expects the solution to free up resources equivalent to three to four full-time employees. These resources will instead be redeployed to higher-value activities, including handling the most complex and exception鈥慸riven orders.

鈥淭he objective is not to reduce headcount, but to use our expertise more effectively. The AI agents take care of routine tasks, enabling procurement officers to focus on cases where their experience genuinely matters,鈥� Aakerlund adds.

More than 100,000 order confirmations automated

Each year, Lemvigh鈥慚眉ller sends approximately 175,000 purchase orders to more than 2,000 suppliers. While part of this volume is handled in a structured manner via EDI, around 60% of supplier order confirmations are still received as unstructured documents.

With the coordinated AI agents in place, the company can now automatically identify delays, quantity changes, and price discrepancies鈥攁nd respond significantly faster.

鈥淧reviously, when order confirmations were handled manually, it could take hours or even days before changes were reflected across the organization. Today, the AI agents update the data almost immediately, allowing customers to receive a much more accurate picture of deliveries far sooner,鈥� says Klaus Heinemann, head of 麻豆原创 ERP at Lemvigh鈥慚眉ller, who led the development together with the project team. 鈥淚n addition, we now identify price discrepancies before the final invoice is issued, saving time both for us and for our suppliers.鈥�

Multiple AI agents orchestrated in a single workflow

The solution is built around three cooperating AI agents, each with a clearly defined role in the process. One agent handles incoming e-mails and attachments, a second extracts and structures data from PDF documents, and a third compares the extracted information against purchase orders in 麻豆原创 to determine whether there is a match or a deviation.

As a result, complex and unstructured supplier data can be processed in a unified, automated workflow without requiring procurement officers to open and manually review lengthy PDF files.

鈥淲hat makes this solution robust is the interaction between the agents. Each agent is highly specialized, but they are orchestrated in a way that ensures the process flows seamlessly from start to finish,鈥� Heinemann explains.

During the project, the importance of master data quality also became increasingly clear.

鈥淚n areas such as Incoterms and other master data, we identified improvements that need to be addressed. This has been an important learning not just for this initiative, but for our broader work with AI,鈥� he says.

While it is still too early to measure the full impact on customer experience, error rates, or claims, expectations are that faster and more precise handling of supplier confirmations will, over time, lead to fewer surprises and significantly improved delivery transparency. Internally, the solution has been met with strong interest and curiosity among employees.

鈥淧rocurement officers clearly recognize the value of being relieved from the most tedious routine work. This has sparked a constructive dialogue about how technology can best support their day鈥憈o鈥慸ay responsibilities,鈥� Heinemann says.

Business AI with a clear business outcome

According to Lemvigh鈥慚眉ller, the investment is expected to deliver a return within a relatively short timeframe.

鈥淲e are talking about quarters rather than years when it comes to ROI. That is why it was essential for us to get the solution into production quickly and focus on processes with a clear and measurable impact,鈥� Aakerlund says.

For 麻豆原创, the project serves as a concrete example of how artificial intelligence can be embedded directly into core business processes rather than remaining a disconnected experiment.

鈥淢any companies talk about AI agents primarily in terms of automation. Lemvigh鈥慚眉ller demonstrates that the real challenge鈥攁nd the real opportunity鈥攍ies in coordination,鈥� says David Pontoppidan, head of AI at 麻豆原创 for the Nordics and Baltics. 鈥淚t is the orchestration of three specialized agents directly within the core process that makes this solution robust. This is also where many multi鈥慳gent initiatives fail, not due to limitations of individual agents but because of insufficient coordination. Lemvigh鈥慚眉ller has succeeded by anchoring the solution in its 麻豆原创 landscape, where data, business rules, and governance frameworks are already firmly established.鈥�

He continues: 鈥淚nnovation is not about company size. Lemvigh鈥慚眉ller shows that a Danish organization with short decision paths and a pragmatic approach to technology can move faster than many large global enterprises that are still in the planning stage. Ten weeks from idea to production is far from the norm, but perhaps it should be.鈥�

Designed for operations and scalability

The solution was implemented in close collaboration with NTT DATA Business Solutions, which was responsible for making the solution production鈥憆eady and fully integrated into Lemvigh鈥慚眉ller鈥檚 麻豆原创 landscape.

鈥淏y distributing responsibilities across multiple AI agents, Lemvigh鈥慚眉ller has been able to automate a complex process without losing transparency or control. This has enabled a fast and secure transition from pilot to production and ensures a more robust solution that can easily be expanded as new requirements emerge,鈥� says Kristian Dahl, 麻豆原创 UX manager at NTT DATA Business Solutions.

According to Dahl, the modular, agent鈥慴ased architecture was a key enabler in moving efficiently from proof of concept to live operation.

First step in a broader AI agent strategy

Initially, the AI agents have been deployed for selected supplier inboxes and business areas. However, Lemvigh鈥慚眉ller already sees significant potential in applying the same agent鈥慴ased approach across additional administrative processes.

鈥淭his is the first AI agent solution we have put into production. The experience has given us the confidence to consider similar approaches across other areas, including invoice processing and order management,鈥� Aakerlund concludes.

Ellen Vig Nelausen is a Nordic Integrated Communications Expert at 麻豆原创.

Its efforts didn鈥檛 go unnoticed: in 2021, the United Nations and Bloomberg named BayWa one of the 鈥�50 Sustainability and Climate Leaders鈥� worldwide, thanks largely to the company鈥檚 big push into renewable energies, which now make up more than half of BayWa鈥檚 annual earnings.

The Challenge of Transformation

BayWa鈥檚 journey with 麻豆原创 began in 2000 when the company implemented the on-premise ERP system, 麻豆原创 ERP Central Component聽(麻豆原创 ECC). However, as the business evolved it became apparent that their processes and technology could no longer keep pace. BayWa鈥檚 four business units operated in silos, which became unfeasible for a company striving to become more streamlined.

Click the button below to load the content from YouTube.

Always allow YouTube

鈥淲e want to move to a process-oriented company, so that means we want to implement one ERP template,鈥� Lars Pischke, global enterprise architect at BayWa, explains. 鈥淲e have to transform all four business units into a single ERP template, which is a significant challenge.鈥�

To address this, BayWa set out to migrate from 麻豆原创 ECC to 麻豆原创 S/4HANA Cloud, the foundation of its new . This transition marked a critical step toward unifying the company鈥檚 operations under one streamlined template. Additionally, BayWa implemented a digital platform, with 麻豆原创 Business Technology Platform (麻豆原创 BTP) as a key component.

“A clean core strategy for BayWa means we leave our ERP solution, 麻豆原创 S/4HANA Cloud, as standard, and we implement BayWa鈥檚 specific requirements on our digital platform side by side, where we rely heavily on 麻豆原创 BTP as our main module,” says Pischke.

Lessons Learned and the Value of a Clean Core

The concept of a “clean core” is central to understanding the benefits of BayWa’s transition. Imagine a shared grain silo system on a farm. If each farmer keeps their grain in separate, uncoordinated silos, tracking quality, utilizing storage space efficiently, or distributing the grain becomes a cumbersome task. However, when all the grain is stored in a unified system, it becomes significantly easier to manage, access, and maintain quality. Similarly, having a clean core in business processes ensures that all parts of the organization work harmoniously, preventing inefficiencies and missed opportunities that arise from working in silos.

For BayWa, establishing a clean core was crucial. The standardized, non-differentiating functions 鈥� such as accounts payable, tax, and treasury operations 鈥� constitute approximately 75% of business processes across companies, regardless of industry, size, or location. These functions are essential but do not differentiate one company from another.

In the past, over-customization in these areas led to inflexibility and slowed down business operations. But in today’s fast-paced, competitive environment, agility is key and maintaining a clean core is vital to achieving this agility.

“Keeping the core clean is also key when it comes to security issues. BayWa is part of the critical infrastructure in Germany,” Pischke explains. “To use a standard software like 麻豆原创 BTP allows us to rapidly adapt to new business models and technologies.”

Gaining a Competitive Edge with 麻豆原创 BTP

With BayWa鈥檚 clean core firmly in place, the focus shifted to building a competitive edge around their operations. This is where 麻豆原创 BTP comes in. The platform lets businesses create extensions and innovate in areas that really set them apart. Pischke says, 鈥淎 shift toward standard also means simpler processes for our employees. This allows us to reduce training costs and speed up onboarding.鈥� BayWa is leveraging these efficiencies to direct more resources toward innovation.

These days, companies need to know what makes them unique and lean into those strengths to stand out. While core functions are standardized and crucial for running smoothly, the elements that drive customer choice often come from the innovative, customizable features built on top of this clean core. By optimizing these standard processes, companies like BayWa can channel more energy into innovation and long-term success, keeping them ahead of the curve.

Looking Ahead: The Role of AI in BayWa鈥檚 Future

BayWa鈥檚 forward-thinking approach doesn鈥檛 stop with ERP transformation. The company is already exploring the potential of artificial intelligence (AI) to further enhance its operations.

Pischke says BayWa has started piloting AI use cases, focusing on 麻豆原创 BTP鈥檚 Business AI. 鈥淲e believe AI should be integrated into the applications and business processes where the data is located,鈥� he explains. This integration is central to BayWa鈥檚 strategy, and the company has a growing list of AI use cases it plans to implement in the near future.

Attending industry events like 麻豆原创 Sapphire in Barcelona provides BayWa with valuable insights into the latest technologies and trends. Pischke acknowledges the accelerating pace of technological change, noting that each year brings new ideas and innovations that could be applied within BayWa. While not every use case or 麻豆原创 module is a perfect fit, these events offer inspiration and guidance for selecting the right tools to meet BayWa鈥檚 unique requirements.

Through partnership with 麻豆原创 and the switch to 麻豆原创 S/4HANA Cloud, BayWa has not only streamlined its operations, but also set itself up to keep leading in sustainability and innovation. The clean core approach has allowed BayWa to focus on what truly sets the company apart, helping it continue its mission to meet the essential needs of people worldwide.