Process mining provides companies with fertile ground for AI deployment

In order to generate real value, artificial intelligence (AI) must first understand how business operations work. At the annual Celosphere event, software supplier Celonis explains how it can help companies achieve this. Various customers, including Mercedes-Benz, Andritz, Thyssenkrupp and Vinmar, provide practical examples. ”This approach not only changes processes, but also corporate culture.”

By Marcel te Lindert

In his opening speech at Celosphere, Alexander Rinke (photo) cites a study by IDC. It states that only 11 per cent of companies benefit from all their investments in AI. “Isn’t that strange?” the co-founder and co-CEO of Celonis asks the audience. “It proves that investments in AI are not being recouped. Not yet, at least. And that’s something we all sense, really.”

Software company Celonis has its origins in process mining, but at its annual customer event in Munich, it only talks about ‘enterprise AI’. By this, Rinke means not only the latest innovations in generative and agentic AI, but also algorithms, machine learning and deep learning. ”Enterprise AI is the driving force behind recommendations, predictions, co-pilots, agents, workflows and smart apps. If we succeed in feeding our processes with enterprise AI, it could radically change our business operations.”

The right context

If companies want to successfully apply enterprise AI, they must meet three preconditions. ”First of all, AI must understand how the business works. It needs context,” says Rinke. Second, AI must be deployed strategically in the right areas of the business, and third, it must work together with all the other systems you have.”

According to Rinke, Celonis’ software can provide the right context. Or, as the software supplier likes to say: no artificial intelligence without process intelligence. “That context is about how the processes work in practice. It’s not just locked up in systems, but also in process diagrams, IT architecture, log files, documents, e-mails and other things. Think also of all the manual actions that employees perform to copy something from one system to another.”

Process intelligence graph

Celonis uses process mining to find out how processes work in practice. This is almost always different from how companies think their processes work, let alone how they were originally designed. “We use data from the systems and enrich it with company-specific information, rules, performance indicators and other content to build a digital twin of the operation,” says Rinke, who uses the term ‘process intelligence graph’ for this.

The process intelligence graph, or PIG for short, enables companies to meet the other two preconditions as well. Analysis of the PIG reveals bottlenecks in business operations or opportunities for optimisation. These are the areas where enterprise AI can offer a solution. “Our platform is completely open, so companies can design and deploy their own AI solutions,” says Rinke.

Bottlenecks due to chip shortage

During Celosphere, companies from the German automotive industry in particular talk about how they use process mining. One of them is Mercedes-Benz, which first turned to Celonis during the previous chip crisis. “We had data in different systems, but we weren’t able to link them together. We managed to do so with Celonis’ software, which gave us insight into the real bottlenecks caused by the chip shortage,” says Jörg Burzer, chief technology officer at Mercedes-Benz.

Process mining is now widely used by the Stuttgart-based car manufacturer. It has led to fewer delays in car deliveries, fewer production errors and faster parts logistics. ”The process intelligence graph we have created enables us to learn from our own processes. This approach not only brings about changes in processes, but also changes in our corporate culture,” says Burzer.

Automatically calculate your carbon footprint

Machine manufacturer Andritz uses an app that calculates CO2 emissions based on the process intelligence graph. The CO2 app uses transaction data from the supply chain and enriches it with emissions data. Based on a machine’s bill of materials, the app is able to link all purchase orders, production orders and transport orders with associated emissions to that machine. With just a few clicks, it is clear which supplier, which transport route or which step in the production process causes the most emissions.

“This makes it easy for everyone within the company to calculate the total CO2 footprint. Take our sales manager who is working on a quote for a customer. He gets an accurate answer based on current data,” says Marco Mansoner, data scientist at Andritz. “We can calculate the footprint before the machine is built and recalculate it afterwards when we have received additional data from suppliers.”

Linking planning systems

The Thyssenkrupp site in Rasselstein uses Celonis software to gain more control over incoming goods flows. The company manufactures steel packaging in Rüsselsheim, using steel from the steelworks in Duisburg. ”We often only knew what Duisburg was delivering when we opened the doors of the lorry. To avoid bottlenecks in production, we needed more insight into deliveries,” says Tanja Lommel, Vice President Sales Planning & Logistics at Thyssenkrupp.

With the help of Celonis software, the steel producer has succeeded in linking the production schedules of Rasselstein and Duisburg. The people in Rasselstein can see what Duisburg is delivering and link that to their own steel requirements. This makes it possible to predict shortages at an early stage. With the help of generative AI, this information is presented in understandable language. “We can now adjust the priorities in the Duisburg system to prevent shortages,” says Mario Kossmann, vice president of digital solutions at the Rasselstein site.

Automating order-to-cash

It is not only manufacturing companies that tell their stories on Celopshere. American chemical distributor Vinmar explains how Celonis’ process intelligence graph helps to manage its complex supply chain. “Until recently, booking sea freight, for example, was a manual activity,” says CEO Vishal Goradia. “Now we use AI to compare all potential carriers. We now get a top 3 with the best options that meet our requirements. Once our planner has made a choice, Celonis ensures that the booking is carried out.”

Vinmar is also working on automating the entire order-to-cash process with the help of AI. Goradia: “This allows our employees to focus on solving problems and coordinating tasks. As a result, we can offer our customers a better experience.”

The post Process mining provides companies with fertile ground for AI deployment appeared first on Supply Chain Movement.