Within the project we will explore new challenges and limits at the interface of data science, machine learning and intelligent interface design for the financial sector. The long-term idea/objective here is to develop a (semi-)autonomous intelligent platform (called the "virtual self") for intelligent data management, retrieval and analysis. The "virtual self" is similar to a multi-agent system that learns and acts autonomously and at the same time acts as a supporting tool for the responsible data owner (who has the "last word"). Against this background, the "virtual self" can solve problems whose solution is time-consuming or not apparent to the data owner by incorporating methodological, functional and procedural approaches, algorithmic searches and various learning concepts. Key components of the "virtual self" are its explanatory and comprehensible character, its adaptive structure and its ability to share transparency through communication with users and/or other machines.