This study examines user behavior in using AI tools in solving business cases, applying the Technology Acceptance Model (TAM) and the Fogg model to analyze interaction patterns. Participants were categorized based on collaboration and exploitation levels during task execution. Results indicate that TAM effectively represents general usage experience, while the Fogg model is better in charaterizing specific usage behaviors. Excessive reliance on GPT’s perceived reliability correlated with exploitative usage, while balanced perceptions of efficiency and reliability resulted in collaborative behaviour. Findings reveal differing relationships among items and two behavior types. Despite the exploratory nature and small sample size, this research highlights the complementary value of TAM and Fogg model in understanding GPT interactions. Findings inform the design of AI systems to optimize collaborative potential while mitigating exploitative tendencies.