A large German plant manufacturer for the food & beverage industry receives around 1000 supplier invoices each month. Robotic Process Automation will accelerate the invoice checking process and relieve the purchasing team of this time-consuming workload. Before the software robots start their work, an intelligent semantic analysis prepares the workplace for these digital colleagues.

The construction of end-to-end production lines for the food & beverage industry requires a huge number of individual components and materials, which a German machine and plant manufacturer procures from a large distributed network of suppliers worldwide. This causes a lot of work for the company purchasing department: 1000 times a month, supplier invoices end up in the mailboxes of the clerks; 1000 times, the employees have to check whether the value-added tax is shown correctly in the invoices; 1000 times, the result must be documented with additional information in SAP.

Automate? Yes, but…

Now normally, software robots love structured data with defined categories and clearly assigned values that they can process by column and row. However, the monthly invoices of international supplier companies are typically received as PDFs or JPGs, meaning the relevant data is held within image information, and simple copy-paste transfer to SAP is not possible. In addition, the languages and layouts of the documents vary widely. In short, supplier invoices flood the purchasing department month after month with a lot of unstructured information instead of neatly arranged data. "I can't work this way!" a software robot would complain if it could. "Please clean up first."  

Semantically sorted: structured data for digital co-workers

So how do you filter precisely the data that your digital colleagues can process properly, from the tsunami of incoming documents every month? The IT managers of the plant manufacturer rely on a combined solution from T-Systems and Servicetrace GmbH: with Semasuite, Germany's largest service provider for information and communication technology offers a sophisticated framework for the evaluation and interpretation of natural language texts, and XceleratorOne (X1), the specialists for software robotics deliver the matching platform for end-to-end process automation.

Both technologies integrate seamlessly in this use case: the first step is for Semasuite to perform a semantic analysis of the supplier invoices received by the purchasing department and extract the relevant VAT information. Semasuite transfers this robot-ready data to the XceleratorOne RPA platform for subsequent automated processing: The X1-Bots mail the extracted VAT information to the relevant purchasing staff and maintain all relevant invoice data in SAP.


T-Systems Semasuite X1en


Business Case: Semasuite extracts structured data from the incoming supplier invoices, enabling the X1-Bot to automate invoice verification.

Top Trend: combining RPA smart

RPA is a core technology of digital transformation. The range of possible use cases is considerably broader if additional digital technologies such as natural speech processing are connected

says Enrico Jorgel, project manager at T-Systems Multimedia Solutions, describing the advantages of this combined automation solution. "Semasuite and XceleratorOne offer an integrated solution for data structuring and process automation from a single source."

And customers get a true measure of success: today, the purchasing department of the machine and plant manufacturer is delighted to have saved around 350 working hours every month. This adds up to 4,200 hours per year – extra time for value-adding activities and strategic projects that transform the purchasing department from a simple administrative function to a driver for successful change in the company.