[MA/PA] Augmenting first-level support service with LLM-based conversational agents


We are excited to announce a master thesis opportunity in the field of enhancing first-level support services through LLM-based conversational agents (Knote et al., 2019). In this thesis, students will explore the possibilities of leveraging LLMs, such as models by OpenAI, the models by Google, the models behind Claude, or open source models like Bloom to improve the onboarding process and first-level support for users of newly implemented software in organizations (Huang & Rust, 2018). This research aims to enhance the user experience and streamline support interactions by harnessing the power of cutting-edge language models. Three subject areas are available:

  1. Backend development: Developing language-model-based solutions with frameworks, such as LangChain.
  2. Frontend development: Developing customer-centric user interfaces for conversational agents.
  3. Business application: Testing the developed chatbots as conversational agents in first-level support scenarios.

Problem Statement

The introduction of new software in organizations often requires employees to undergo onboarding and seek first-level support to learn and adapt to the system. Traditional support methods may have limitations, leading to potential inefficiencies and user frustrations. Addressing this gap, this thesis project aims to investigate how LLM-based conversational agents can revolutionize first-level support services (Feine et al., 2020), providing users with interactive and dynamic assistance during onboarding and beyond.

Subject of the Thesis and Directions for Research

The core objective of this master thesis is to explore the capabilities and effectiveness of LLM-based conversational agents in augmenting first-level support services. Students will embark on a research journey and may focus on some of the following key directions:

  1. Evaluating Effectiveness: Assessing the performance of LLM-based conversational agents in providing onboarding assistance and answering first-level support questions, comparing it to traditional support methods.
  2. User Satisfaction and Support Efficiency: Investigating the impact of LLM-based conversational agents on user satisfaction and the overall efficiency of support services, with a focus on metrics such as response time and issue resolution.
  3. Training and Fine-Tuning: Exploring different training and fine-tuning approaches for LLMs to optimize their performance specifically in the domain of software onboarding and support, considering the unique requirements and nuances of this context.
  4. Scalability and Practical Implementation: Examining the practical aspects and scalability of implementing LLM-based conversational agents within an organizational support infrastructure, including considerations of computational resources and integration challenges.
  5. Ethical Considerations: Exploring potential ethical considerations associated with LLM-based support systems, such as data privacy, security, and biases, and proposing approaches to address these concerns.


To pursue subject area 1, students have to work with several NLP-related Python libraries and APIs.

To pursue subject area 2, students have to have experience in user interface design. Initial experience with NLP is helpful.

To pursue subject area 2, students are not required to have programming experience, though it is helpful.

Data collection will be essential, including user inquiries, support logs, and user feedback, to evaluate the performance and impact of LLM-based conversational agents as chatbots. The thesis may involve experimental design, development, and integration of the conversational agent, as well as addressing ethical considerations throughout the research process.

We encourage motivated and enthusiastic students especially in the field of information systems with interests in computer science, artificial intelligence, or related fields to apply. This is an excellent opportunity to contribute to cutting-edge research and make a real-world impact in the field of software onboarding and support.


The thesis and application material have to be submitted in English.

Call for action

Julius Kirschbaum, M. Sc.

Research Associate and Doctoral Student

School of Business, Economics and Society
Chair of Information Systems I, Innovation and Value Creation (Prof. Dr. Möslein)

Please apply to Julius Kirschbaum following the guidelines for thesis applications on our chair’s website.

IMPORTANT: Read this short handout on the thesis process: Student Theses – Process (09.05.23)


Feine, J. et al. 2020. “Designing Interactive Chatbot Development Systems,” in Proceedings of the 41st International Conference on Information Systems (ICIS), AIS Electronic Library (AISeL), pp. 0–17.

Huang, M. H., and Rust, R. T. 2018. “Artificial Intelligence in Service,” Journal of Service Research (21:2), pp. 155–172. (https://doi.org/10.1177/1094670517752459).

Knote, R. et al. 2019. “Classifying Smart Personal Assistants: An Empirical Cluster Analysis,” Proceedings of the Annual Hawaii International Conference on System Sciences (2019-January), pp. 2024–2033. (https://doi.org/10.24251/hicss.2019.245).