AI Governance Model

AI Governance Framework


For instance, Lemonade is an insurance company that leverages the capabilities of conversational AI to revolutionize the traditional insurance model, streamlining the customer experience and offering efficient communication. Lemonade enhances customer interactions and simplifies complex insurance processes by employing conversational AI as a central and dominating customer channel across touchpoints. To do so, it employs a software-based agent capable of participating in natural, human-like user dialogue. The agent is engineered to comprehend, interpret, and produce responses that closely mimic human communication, resulting in more organic and intuitive interactions compared to a traditional website. This elevates the user experience and promotes more effective communication between Lemonade and its customers. As a result, Lemonade can offer instant support, simplify the purchase of insurance policies, and settle claims rapidly, thus setting new standards for customer experience and efficiency in the insurance landscape. However, conversational AI is also disrupting other industries. For example, Emirates Vacation integrated a conversational AI solution within its display ads, boosting engagement rates by 87% during a 30-day test. Likewise, Madison Reed’s Madi, a computer vision-enabled chatbot, increased engagement by 400% and achieved a 21% click-through rate.

Despite such success stories, numerous large-scale conversational AI projects have either failed or encountered severe challenges during their initial stage. Thus, it becomes clear that successfully implementing and managing conversational AI is not as easy as copying another company’s strategy. For example, Microsoft launched Tay, an AI-driven chatbot, in March 2016, only to decommission it within 24 hours due to the system’s dissemination of alarming political views instigated by user manipulation. Another notable event involved Facebook’s M assistant, launched in August 2015 and failing to adequately address approximately 70% of user inquiries. Consequently, Facebook terminated the M assistant in January 2018. In addition, in February 2023, Bing, owned by Microsoft, which integrated Open AI’s GPT4 in its search engine, experienced several glitches, causing Bing to propagate extreme views and suggesting to one user that he should end his marriage. Thus, the primary challenges in implementing conversational AI consist of, among others, establishing a clear vision of how the company will utilize the conversational AI solution to add value, ensuring user acceptance, enforcing adequate security, compliance, and governance measures, and maintaining a team dedicated to the continuous improvement of the solution.

Objective of the thesis

This thesis aims to develop an AI Governance Model for incumbents in the Swiss financial industry. A wide variety of data has already been collected, placing you in a favorable position to bypass the data collection phase and concentrate directly on data analysis. It should be noted that the data consists of anonymized interviews.

Your Profile

  • Have a keen interest in Conversational AI and the financial industry.
  • Be capable of working independently, with a proactive and problem-solving attitude.
  • Possess strong communication skills and be willing to collaborate with a research team.
  • Demonstrate proficiency in scientific writing in English.

We offer

Close supervision by experienced researchers in the fields of Information Systems and Management.

Regular review meetings and feedback sessions to support your progress and development.

A flexible start date, with a completion timeline of approximately 6 months.


A CV outlining relevant experience and skills.




Forschungsgruppe Prof. Dr. Ivo Blohm


Kevin Schmitt

Zum Detail