Supply chain insights with just one click: ALTANA pools knowledge in graph database built on Microsoft Azure

The challenge: Fragmented knowledge spread across 35 ERP systems

When the purchasing team at ALTANA go into meetings with their suppliers, it’s crucial that they have all relevant information such as supply volumes and other benchmark data at their fingertips. The same goes for purchasing management when company-wide data must be compiled to generate overviews of suppliers or price development. These overviews are the basis for optimizing both the business and its products—for instance, when the company is looking to exchange one raw material for another with a smaller carbon footprint.

But preparing for these meetings used to be really complicated: “Our company has a decentralized structure—we use 35 different ERP systems. What’s more, we didn’t have a central repository. This meant that a given supplier might be saved under a variety of IDs,” says Stefan Franke, Head of Procurement Excellence at ALTANA. “So in the past, putting together a company overview of a particular topic took a great deal of time. And once we had the data, it took considerable expert knowledge to interpret it,” Franke continues. His colleague Timo Scheller, Project Manager Sustainability at ALTANA, adds: “It was a question of how to pool our knowledge, establish a uniform understanding of the data, and use synergies efficiently.”

The fact was that fragmented knowledge was clouding the global view of cost transparency, supply chain resilience, and sustainability data. But ensuring a clear view is becoming increasingly important, for example in light of Germany’s Supply Chain Due Diligence Act (LkSG). “Our old solution would’ve left us with the challenge of making data on sustainability, markets, and risks available in multiple tools. We wanted to simplify the situation,” Scheller says.

So ALTANA decided to enlist the help of its long-standing partner PRODYNA to switch as soon as possible to a new platform: EMPOWER. The centerpiece of this solution, which is based on Microsoft Azure and a variety of managed services, is a graph database. “Demolishing silos for good is really about establishing useful data relationships and enabling simple inquiries across domains and departments. Visualizing the data as a graph lets us do just that,” says Maximilian Render, Senior Sales Manager at PRODYNA. The project team had just twelve weeks in which to launch an initial functional platform. “But with the help of Azure DevOps, we soon managed to pick up the pace—it also helped that we were already using Azure within the company,” Franke says.

The solution: A single point of truth through a graph database and Microsoft Azure

Today, the EMPOWER platform is hosted entirely in Azure and is made up of three components: The first is the knowledge graph, which pulls together and combines internal metadata on material categories, raw materials, and suppliers from the 35 ERP systems. The second is the data factory based on Azure SQL Managed Instance, Azure Kubernetes Service, and Azure Blob Storage. This is where the internal data is enriched with external sustainability, market, and risk data. And the third is the user interface that then visualizes the data relationships as a graph—expandable and standardized. “Users see immediately if there are any discrepancies in the data. They can then decide how to map it. The system learns from this process so that the same error gets corrected automatically at each new data import,” Render says. For highly specialized data inquiries such as reports or ad hoc analyses, there is also the option of creating visualizations via a Power BI dashboard.

Merging the data means that algorithms in the data factory can mine other available data to identify supplier risks such as category and country risks, and amend the relevant entry in the graph database. “In the chemical industry, there can be 35 production steps between raw material and finished product, sometimes more,” Scheller says. “Visualizing in graphs brings more transparency to our complex purchasing data. Colleagues can see immediately whether alternative materials are available should a supplier drop out, and how this will affect the price of the raw material portfolio. This provides us with valuable insights into the resilience of our supply chains—and with a single click, too.” In the data factory, artificial intelligence (AI) models will in the future be able to use current raw material prices and market data to predict how prices will develop. The initial test runs show promising results: when making these predictions, the AI was off by just 5 percent. ALTANA will therefore be rolling out this kind of AI modeling for the entire purchasing team by the end of 2023.

EMPOWER’s interface and integration options make the platform extremely flexible and expandable. “Say there’s a shift in legal or internal requirements that means we want to change the way we make sustainability data available, all we’ll have to do is adapt the data model. Thanks to EMPOWER and Azure, we can respond and grow quickly,” Franke says.

The EMPOWER platform has made it much easier for the purchasing department to see what’s going on company-wide. “Despite our decentralized structure, we now have a simple and transparent way of sharing data and knowledge. And since identifying data relationships no longer requires IT expertise, everyone benefits,” Scheller says.

In the future, ALTANA’s priority will be to further enrich its data and to take data-driven decisions and measures that make the company even more resilient. “We currently have around 50,000 raw materials in the system. Generative AI will allow us to add information relating to classification or production quicker to these entries and thus expand our store of knowledge,” Franke says.