Large Language Models (LLMs) are increasingly used to support managerial decision making by generating analyses, recommendations, and strategic alternatives. While generative AI has demonstrated substantial productivity gains in knowledge work, it remains susceptible to hallucinations, incomplete reasoning, and overconfident but incorrect outputs. As a result, managers cannot simply accept AI-generated recommendations at face value but must evaluate, verify, and adapt them before incorporating them into organizational decisions.

Existing research has primarily focused on trust in AI or on the quality of AI explanations. However, much less is known about what managers actually do after receiving AI-generated advice. They may seek external information, consult colleagues, challenge assumptions, ask follow-up questions, or triangulate recommendations with additional evidence. These verification behaviors represent a critical yet underexplored step between AI-generated knowledge and organizational decision making.

Understanding these verification strategies is essential for designing AI decision support systems that foster appropriate reliance rather than overreliance or unnecessary skepticism.

The thesis is supposed to answer the research question: "How do managers verify AI-generated recommendations before incorporating them into organizational decisions, and which verification strategies emerge across different decision contexts?”

If you are interested in this theses, feel free to reach out to Tobias Kotzian (tobias.kotzian@unisg.ch).