Final theses in the field of Agentic AI - Soloplan GmbH

In the summer semester of 2026, the Chair of Intelligent Systems and Services will start an exciting cooperation with Soloplan GmbH, which will enable students to carry out their final theses and research projects in cooperation with an internationally active company and develop innovative solutions for the logistics industry.

Soloplan GmbH is a family-run, medium-sized software company with its headquarters in Kempten in the Allgäu region and other locations in Dresden, Hamburg, Poland, Spain and South Africa. Since its foundation in 1991, Soloplan has been one of the leading providers of logistics and transportation management software.

The company's core product is the "CarLo" software , a solution for planning, controlling and optimizing transport processes. It is used worldwide by customers from the forwarding, industry and logistics sectors to make processes more efficient and transparent. With around 260 employees and more than 1,800 customers in over 50 countries, Soloplan stands for innovative IT solutions that make an important contribution to the digitalization of the logistics industry. The company is also actively involved in cooperation with universities, particularly in the areas of research and development. Soloplan GmbH

Topics for final theses in the summer semester 2026
In cooperation with Soloplan GmbH, the following three topics are offered, which are aimed at innovative technologies and approaches in the field of "CarLo" transportation management software

• Security Mechanisms for Agentic AI in Enterprise Systems: Prompt Injection and Tool Abuse Mitigation
  The aim of this work is to investigate how agentic AI functions in the "CarLo" system can be protected against threats such as prompt injection, tool abuse, data exfiltration and malicious instructions. A security architecture with context separation, policy control, validation middleware and auditable logging for enterprise use is to be developed and evaluated.
• Coordination Strategies for Multi-Agent Systems in Transport Management This work deals with the question of how several specialized AI agents within the "CarLo" ecosystem can work together to support complex logistical decisions. The aim is to design coordination strategies for areas such as planning, cost, capacity, risk and compliance and to evaluate their suitability for future scenarios in transportation management.
• Feedback-Driven Learning for Adaptive Agentic Systems in Logistics
  The aim of the work is to analyze how agentic AI systems in the "CarLo" ecosystem can be improved through structured feedback loops. Explicit feedback, implicit corrections and operational results are to be recorded in order to optimize future agent decisions and enable controlled, reproducible updates.

Application and further information
Students who are interested in one of the above-mentioned topics can contact the chair immediately. The projects offer an excellent opportunity to combine innovative technologies and practice-oriented research and to contribute to the further development of the logistics software "CarLo". Theses - Intelligent systems and services