We look for motivated students who would like to complete their Master’s thesis in a collaboration between the Institute of Information Systems (IWI-HSG) at the University of St.Gallen and the German online retailer OTTO (GmbH & Co KG).

Objective
The objective of this thesis is to examine the long term impact over time of the initial implementation of an AI-supported sales forecasting system at OTTO, one of the largest German online stores. For this purpose, different focal points such as purely economic, social-organizational or technical aspects, can be selected. The initial implementation of the AI-supported sales forecasting system can be monitored by conducting quantitative and qualitative analyses. In particular, aspects like “perceived AI performance”, “behavioral intention”, “actual system use” and “actual task performance” shall be operationalized and monitored. To this end, cooperating departments, which use the new AI sales forecasting system within Otto can be determined in joint coordination.

Mode of collaboration
The thesis is written in the practical environment of OTTO. The company headquarters is located in Hamburg, Germany. No personal presence on site is necessary and the research work can mainly take place digitally. During the elaboration you will be accompanied by the internal team, which is responsible for forecasting, and you will also come into exchange with various other departments.

Requirements
Your thesis is an intellectual challenge! It is well-suited for ambitious, communicative, and curious students who are keen to work within a strategic, operational environment. In return, you will gain experience and knowledge in this up-and-coming field of management, and address future challenges of modern business organizations while advancing your knowledge on the management of AI. Basic SQL and programming skills are required to run quantitative analyses.

Application
The nature of the work requires your full attention. The thesis can be started by agreement. Please send your application including cover letter, CV, and current transcript of grades to ­ Tobias Fahse .

Why?
The current rise of online learning environments, such as massive open online courses or learning management systems comes with many new opportunities. Therefore, now is the time to leave behind the one-size-fits all mentality in education and personalize education. The individualization of online trainings (e.g. Cousera courses) enables users to learn exactly their required knowledge, within their own pace and style. This master thesis aims to identify different parameters for personalization in intelligent learning systems.

How:
Master thesis that conducts a systematic literature review using the methodology of Webster and Watson (2002) and Vom Brocke et al. (2009; 2015).

Key-Facts:

• Close supervision, with regular review meetings, feedback discussions, etc.
• Thesis can start immediately, but should be completed within the next ±6 months
• If the work is worth publishing, you will be listed as author

If you are interested, send your CV, transcript of records and a brief description of your motivation to Eva Ritz (). Looking forward to hear from you!

Why:

The OECD estimates that the 4th Industrial Revolution will change 1.1 billion jobs, creating massive skill gaps in organizations’ workforces. Simultaneously, many organizations currently face the challenge of agile transformations. The task of staffing suitable and skilled employees on projects is becoming more important. The use of skill profiles provides the foundation of companies to decide on re-& upskilling measures, identify skill gaps, and match potential project partners (Ritz et al. 2023). Artificial intelligence can use these profiles, for instance, to find and connect employees with similar skills throughout the organizations and break up knowledge silos, and improve project staffing. A permanent diversity concept is called “cognitive diversity”, which is the variety of ways that people think and process information. This master thesis aims to investigate if AI can improve the matching of cognitive diverse project teams.

How:
Master thesis that explains the theoretical concept of “cognitive diversity” and conducts a systematic literature review using the methodology of Webster and Watson (2002) and Vom Brocke et al. (2009; 2015) or conducts an experiment in close supervision.

Key-Facts:

• Close supervision, with regular review meetings, feedback discussions, etc.
• Thesis can start immediately, but should be completed within the next ±6 months
• If the work is worth publishing, you will be listed as author

If you are interested, send your CV, transcript of records and a brief description of your motivation to Eva Ritz (). Looking forward to hear from you!

Many contact centers now use chatbots and voice bots to automate selected customer interactions. However, since these solutions often do not yet work perfectly, they still need to be monitored by a human who may need to intervene in the conversation in some cases. At the same time, there are still many questions about how this type of human-AI collaboration can be designed: How many bots can one employee monitor? How should the conversations be presented so that an employee can intervene when needed? How do you design the handover between humans and machines? In collaboration with a practice partner, this master thesis aims to clarify these questions and develop a dashboard for monitoring bots in the contact center following a human-centric design approach. Specifically, the student will derive design principles based on a literature review and user interviews, implement them in a clickable prototype, and evaluate the dashboard in a user study.

You should

– have a basic understanding of human-centered design methodologies
– are interested in research on Human-AI collaboration (no programming skills needed)
– are passionate about developing and evaluating real-world solutions for practice
– are interested in working together with an industry partner

We offer you close supervision and the opportunity to develop practical as well as theoretical skills in the area of human-AI collaboration. If you are interested, send your CV, transcript of records, and a brief description of your motivation to Kevin Schmitt ().

This qualitative master thesis…

… explores the phenomenon of virtual influencers in virtual world settings. The student should conduct a qualitative study, interviewing online users that are demonstrating regular engagement with (content of) virtual influencers on social media, e.g., a metaverse. The thesis addresses how virtual influencers are defined and perceived by their followers. The thesis aims to elaborate chances and risks of utilizing virtual influencers for marketing purposes, thereby, receiving a condensed image on this type of influencer. This thesis should investigate the follower’s perspective.

Background

By allowing people to interact with one another in immersive, three-dimensional virtual worlds, the metaverse provides opportunities for brands to connect and build relationships with consumers in new ways (Schöbel and Tingelhoff, 2023). A perfect example of this is the rise of virtual influencers (Moustakas et al., 2020).

A virtual influencer can be defined as an embodied digital character – humanlike or not. The character is equipped with its own personality, can be human-controlled or autonomously controlled by an AI, and is made accessible via social media, e.g., the metaverse. Such virtual influencers can impact a vast amount of people. For example, Lil Miquela has more than 7 million followers on different social networks (Instagram, TikTok, Facebook, YouTube) and was voted one of the 25 most influential people on the internet by Time magazine in 2018 alongside personalities such as former President Donald Trump (Time Magazine, 2018; Block & Lovegrove, 2021; Oliviera & Chimenti, 2021). Virtual influencers are also starring in campaigns for global brands such as Prada, Porsche, Unilever, and Ikea, showcasing potential for success as brand ambassadors.

Despite, little is known about the effectiveness of using these characters for marketing purposes. By following a qualitative research approach (interview study), this thesis will contribute to the basic understanding of virtual influencers by identifying potential challenges, opportunities, and the value factors of virtual influencer marketing in the context of social media and the metaverse. In understanding the key factors that compel online users to follow and interact with the digital characters, this thesis broadens the scope of influencer marketing to include the role of virtual influencers. Thereby, is a fundamental piece for researchers and practitioners in the field of both metaverse marketing and influencer marketing.

You should…

• have a basic understanding of metaverse and (virtual) influencers
• have a basic understanding of theories and theorizing
• have a basic understanding of qualitative research methods and content analysis
• be a communicative and open person
• work in an independent and accurate way
• have scientific writing skills in English
• be passionate about the metaverse and how it reshapes business environments

We offer you a close supervision and the opportunity to develop practical, as well as theoretical skills in the area of information systems. If you are interested, send your CV, transcript of records, and a brief description of your motivation to Fabian Tingelhoff (). We are looking forward to hearing from you!

This qualitative master thesis…

… explores the phenomenon of virtual influencers in virtual world settings. The student should conduct a qualitative study, interviewing marketing professionals and influencers that are (thinking about being) active in a metaverse. The thesis addresses how virtual influencers are defined and perceived in practice. The thesis aims to elaborate chances and risks of utilizing virtual influencers for marketing purposes, thereby, receiving a condensed image on this type of influencer. This thesis should investigate the practitioner’s perspective.

Background

By allowing people to interact with one another in immersive, three-dimensional virtual worlds, the metaverse provides opportunities for brands to connect and build relationships with consumers in new ways (Schöbel and Tingelhoff, 2023). A perfect example of this is the rise of virtual influencers (Moustakas et al., 2020).

A virtual influencer can be defined as an embodied digital character – humanlike or not. The character is equipped with its own personality, can be human-controlled or autonomously controlled by an AI, and is made accessible via social media, e.g., the metaverse. Such virtual influencers can impact a vast amount of people. For example, Lil Miquela has more than 7 million followers on different social networks (Instagram, TikTok, Facebook, YouTube) and was voted one of the 25 most influential people on the internet by Time magazine in 2018 alongside personalities such as former President Donald Trump (Time Magazine, 2018; Block & Lovegrove, 2021; Oliviera & Chimenti, 2021). Virtual influencers are also starring in campaigns for global brands such as Prada, Porsche, Unilever, and Ikea, showcasing potential for success as brand ambassadors.

Despite, little is known about the effectiveness of using these characters for marketing purposes. By following a qualitative research approach (interview study), this thesis will contribute to the basic understanding of virtual influencers by identifying potential challenges, opportunities, and the value factors of virtual influencer marketing in the context of social media and the metaverse. In understanding how virtual influencers affect digital marketing communications, this thesis broadens the scope of influencer marketing to include the role of virtual influencers. Thereby, is a fundamental piece for researchers and practitioners in the field of both metaverse marketing and influencer marketing.

You should…

• have a basic understanding of metaverse and (virtual) influencers
• have a basic understanding of theories and theorizing
• have a basic understanding of qualitative research methods and content analysis
• be a communicative and open person
• work in an independent and accurate way
• have scientific writing skills in English
• be passionate about the metaverse and how it reshapes business environments

We offer you a close supervision and the opportunity to develop practical, as well as theoretical skills in the area of information systems. If you are interested, send your CV, transcript of records, and a brief description of your motivation to Fabian Tingelhoff (). We are looking forward to hearing from you!

Do you want a tool that knows it all or many that all know something?

Background:
Many modern systems rely on the integration of different modules (e.g., Alexa skills). Do users prefer if such systems are presented as unitary multitalented agents or is there a benefit to be had in splitting a system up into many experts? Possible benefits include quarantining against generalization of negative impressions caused by submodules of a system or more leeway against perceptions of, e.g., arrogance in self-praise.
A possible, although not mandatory, theoretical entry to hypothesis formation comes from research on social stereotypes. Interaction and social perception are heavily influenced by stereotypes (Snyder 1977). This effect is diluted if stereotype-irrelevant information is presented as well (Dreu 1995). As some positive effects of expertise rely on stereotypes, the construal of conversational agents may suffer if too many skills are categorized under one persona. In choosing this thesis, you will try to inform future research and practice whether and how technological skills should be split or united in their presentation.

How?
First, you will review the relevant literature.
Depending on your findings, you will then carry out an interview series (if you think more exploration is necessary) or a hypothesis testing survey (if you think current research only needs to be extended).

If you are interested in this Bachelor thesis, please send an email to to discuss how to start.

References
Dreu, Carsten K.W. de; Yzerbyt, Vincent Y.; Leyens, Jaques-Philippe (1995): Dilution of Stereotype-Based Cooperation in Mixed-Motive Interdependence. In Journal of Experimental Social Psychology 31 (6), pp. 575–593. DOI: 10.1006/jesp.1995.1026.
Snyder, Mark; Tanke, Elizabeth Decker; Berscheid, Ellen (1977): Social perception and interpersonal behavior: On the self-fulfilling nature of social stereotypes. In Journal of Personality and Social Psychology 35 (9), pp. 656–666. DOI: 10.1037/0022-3514.35.9.656.

What is a chatbot? Do they simulate people or simply conversation? And are all of them the same?

Background:
Virtual agents are often treated as social actors (Nass & Moon, 2000). Therefore, state-of-the art projects in the space of chatbot technology often use insights from anthropomorphism theory (Epley, 2007) to inform their design choices. However, there is considerable variance in how people think about chatbots. This variance is found between distinct kinds of chatbots and between different sets of people.
Skilled conversation (e.g., counselling) is a kind of work that has not been automated to the same degree as less skilled conversation (e.g., customer support). Besides technical challenges, this may also be explained by a certain reluctance in people trusting any such replication effort. A principal factor where such doubt persists is the replication of human-to-human professional relationships in bot-to-human conversation.
There may be multiple pathways to resolve this tension. However, it is yet unclear whether efforts should flow to anthropomorphism, as some believe, or whether chatbots may be framed differently to circumvent the issue, e.g., as tools. In choosing this thesis, you would work to clarify where automation of skilled conversation can achieve the best “bang for buck” and thus contribute to the acceleration of progress in this domain.

How?
Building on existing interviews, you will explore the framing of chatbots as tools relationship versus potential relationships.
Depending on your skills and interests, this work can take different forms.
You may continue the qualitative road and extend our knowledge with a more specialized interview series you develop and carry out.
Or you can prepare a survey-driven experimental approach to inform the decision.

If you are interested, please send an email to to discuss how to start.

References
Epley, Nicholas; Waytz, Adam; Cacioppo, John T. (2007): On seeing human: a three-factor theory of anthropomorphism. In Psychological review 114 (4), pp. 864–886. DOI: 10.1037/0033-295X.114.4.864.
Nass, Clifford; Moon, Youngme (2000): Machines and Mindlessness: Social Responses to Computers. In J Social Isssues 56 (1), pp. 81–103. DOI: 10.1111/0022-4537.00153.

Would you rather give a straight answer than tick dozens of boxes?

Background:
In roughly the last hundred years, questionnaires with rating scales have enjoyed widespread adoption in many domains encompassing everything from basic social science to market research. They are used to collect data for quantitative purposes (Rowley 2014). The methodology has two distinct benefits: questionnaires require little effort to fill in and are straight-forward to analyse.
However, questionnaires come with drawbacks too. For one, scale construction for valid questionnaires is very labour intensive, and many already validated questionnaires are hard to find or must be licensed. Participants are forced to choose between close-ended answers or must rely on abstractions like numbers to relay their actual thoughts or feelings. Maybe most importantly, as easy as it is to fill in a questionnaire, rating long item lists gets tedious.
Recent advances in natural language processing allow it to retain most of what is good with questionnaires without these drawbacks. Transformer-embeddings can be used to model constructs traditionally inferred from rating scale questionnaires with answers to simple questions (Kjell 2022). However, it is yet unclear how these advances can be applied in a way that allows large-scale adoption. In choosing this thesis, you will try to make innovative technology accessible to researchers and companies who want open ended answers they can count on.

How?
Learn how the technology works doing a short narrative literature review.
Define a target demographic and conduct interviews to inform the design of an application embedding the technology.
Build the application (e.g., a website).
Evaluate your work using user feedback.

If you are interested, please send an email to to discuss how to start.

References
Kjell, Oscar N. E.; Sikström, Sverker; Kjell, Katarina; Schwartz, H. Andrew (2022): Natural language analyzed with AI-based transformers predict traditional subjective well-being measures approaching the theoretical upper limits in accuracy. In Scientific reports 12 (1), p. 3918. DOI: 10.1038/s41598-022-07520-w.
Rowley, Jenny (2014): Designing and using research questionnaires. In Management Research Review 37 (3), pp. 308–330. DOI: 10.1108/MRR-02-2013-0027.

Analysing processes and digitalisation in the private equity industry.

Why

After an unprecedented period of growth in the last decade, today private equity funds hold more than 6.3 trillion dollars in assets under management and have become a staple in portfolios of institutional investors ([1] McKinsey, Global Private Markets Review, 2022). As this rise was fuelled by astonishing profitability (it consistently outperformed all other asset classes in the last decade ([1])), research in the last years has mainly focused on explaining the reasons for this outperformance ([2] Kaplan and Schoar, 2003; [3] Harris et Al, 2012; [4] Demaria, 2015). However, few studies look at the processes necessary to achieve this performance. The question becomes a multi-dimensional one when realizing that the processes of a fund are probably not only dependent on its type but also on its investment strategies and degree of technological maturity.

Scope

We at the IWI aim to combine sophisticated academia and practical insights. While building on profound research, your thesis should generate actionable insights with practical applicability. This can be done with a multitude of scientific methods such as desk research, reviewing scientific literature, and conducting hands-on case interviews.
However, to write an exceptional thesis it is necessary for you to find a topic, which excites you and fits your skills and interests. Therefore, the following topics are for inspiration only and I am looking forward to your ideas:

• Analysis of the impact of a private equity fund’s investment strategy on its internal processes and use of technology.
• Use and potential of new technologies in different stages of the private equity investment process.
• Examination of how target companies, general partners and limited partners interact with each other and which structures and technologies they are using for it.

While the questions mentioned above are more suited for qualitative research, I am happy to discuss more quantitatively focused theses as well, if they should better fit your skill profile.

What we expect

• You have an organized, conscientious and result-oriented working attitude
• You want to write a best-in-class master thesis which contributes to state-of-the-art research
• The thesis can start immediately but should be completed within the next ±6 months

What we offer

• Close supervision and support with regular review meetings, feedback sessions and discussions
• Access to experts in the field
• Possibility of making a real impact through publishing in a scientific journal

If you are interested just book a slot with me (https://calendly.com/fabian-karst/master-thesis-consultation) and I am looking forward to your thesis ideas.

Why?

The metaverse has become an interesting phenomenon in the last year and has attracted lots of attention.

However, it is unclear whether our behavior will differ the metaverse from our real-life behavior. In particular, it is interesting to investigate, if we disclose information differently.

Your bachelor’s/master’s thesis will contribute to the understanding of how a metaverse should be designed in the future to ensure data privacy.

What is the metaverse?

Metaverse and its “combination of technological, social, and economic drivers […] is leading to explosive interest.” (Moy and Gadgil 2022, p. 5) While some see Metaverse as the next evolutionary stage of the Internet, others still doubt its value proposition (Stokel-Walker 2021).

Currently, companies are racing to be among the first to provide a metaverse: While Meta, formerly known as Facebook, plans to invest about $10 billion annually for the next 8 years, Microsoft just spend $68 billion in an acquisition to realize their version of a metaverse. While a platform’s success depends on the quantity and quality of members (also known as complementors) (Parker et al. 2016), clearly, these companies must expect to attract more and better complementors to their metaverse as their opponents. For a complementor, the platform’s facilitation-capabilities in its value creation are of utmost importance in the decision to join a platform. (Gawer 2021; Hein et al. 2020) Therefore, to understand metaverse success, it is imperative to investigate value facilitation mechanisms and how metaverse providers are planning to implement those.

How.

– We plan to conduct an experiment on a metaversion in collaboration with the University of Rotterdam to analyze how people disclose information.

You should…

• have a basic understanding of metaverses and their underlying technologies
• have a basic understanding of theories and theorizing
• have a basic understanding of quantitative research methods and an interest in experimental design
• have scientific writing skills in English

Key Factors:

• Collaboration will take place in close supervision, with regular review meetings (biweekly), feedback sessions, etc.
• International team in cooperation with the University of Rotterdam.
• Work can be started immediately, while the experiment is planned for the autum semester
• Possibility of making a real impact through publishing in a scientific journal

If you are interested, please send a mail to Edona Elshan and Dr. Roman Rietsche. We look forward hearing from you!

Unfortunately…

… there are currently no published master thesis on ‘metaverse’ available. However, should you be interested and passionate about ‘metaverse’ and are keen to write your thesis with us, I still want to encourage you to reach out! If you and your topic-idea convince us, there is no reason why a thesis with us shouldn’t be possible! However, please already include a concrete research idea and methodology in your initial request.

Background

Technologic development is changing the World-Wide-Web as we know it. The current version of the “Internet of Information”, used primarily to seemingly exchange information, increasingly evolves to an “Internet of Value”, also called web3, which empowers people to transact items of value with one another. (Tapscott and Tapscott 2017) Indeed, there are key technologies to permit value creation and value exchange in an online setting like cryptocurrencies, NFTs, native coins, and self-sovereign interoperable identities. However, there needs to be a platform so that people can experience web3 and interact with one another. A platform that makes web3 ‘tangible’ is metaverse.