Robotic Process Automation (RPA) refers to software robots that perform tasks exactly as human clerks do.


Software robots offer more mature, scalable, and reliable implementation for repetitive processes, and the employee is free to focus on value-added tasks.


The goal of the research project is an intelligent software system that captures process knowledge, processes it, learns it independently and executes it automatically. In this way, a bridge is built between automated process acquisition and the transfer to a software robot.

Your contact

Portrait Mitarbeiter Sassan Torabi-Goudarzi

Sassan Torabi-Goudarzi

eMail: Sassan.Torabi-Goudarzi@aws-institut.de
Phone: +49 681 96777 629

In a nutshell


As a project funded by the German Federal Ministry of Education and Research (BMBF), KI.RPA aims to develop an intelligent software system that captures and processes process knowledge, learns the underlying process independently and executes it automatically. This saves process costs to a considerable extent, as routine tasks no longer have to be processed by employees, who thus have more time for demanding, creative tasks. Automated data entry and a self-learning system strengthen SMEs in particular through intelligent solutions and thus achieve an overall economic benefit.


The screen actions of all employees involved, such as text entries, mouse clicks or program calls, are recorded. These are consolidated into a business process. The recording program, which runs in the background, records, identifies and anonymizes the relevant process actions. To train the software robot on this basis, a suitable data model is used to mirror the captured process information. In this way, a bridge is built between the automated process recording and the transfer to a software robot.

initial situation

Künstliche Intelligenz Robotic Process Automation

Ever shorter development cycles, cost savings and higher demands on service or product quality present companies with the challenge of using their resources more efficiently with the help of digitalization.

Robotic Process Automation (RPA) mimics human interaction with user interfaces (e.g., on screen or keyboard) and thus offers a way to free human labor from monotonous tasks and use it for value-added tasks. Compared to human input, software robots offer a more mature, scalable and reliable approach to repetitive processes. However, today’s process automation solutions represent a significant manual effort, as business processes must either be captured and neatly documented or elaborated by a domain expert to then manually configure the software robot.


650% in 3 years

montone tasks:

80% of working time

The August-Wilhelm Scheer Institute develops 2 basic components:

Process Model. Relevant process information is automated using artificial intelligence, identified and captured through desktop activity mining (DAM).


Process Configuration Component. The development includes the implementation of a suitable data model to represent the captured process information, the algorithm and the user interface for dynamic interaction with the process model.

press releases

This research project has been completed.



The KI.RPA project is funded by the German Federal Ministry of Education and Research.