Adaptive Recruitment Interfaces Powered by LLMs

Sammanfattning

This thesis explores how generative AI integrated into the front end of a recruitment website affects user behavior and perceived user experience. Two versions of the website were implemented and compared using A/B testing with a test group consisting mainly of students. Layout A served as the baseline version, while Layout B used generative AI to create personalized job summaries and visual elements based on each candidate’s CV. The results from the user tests showed that the generative AI version had a positive effect on several engagement-related metrics and aspects of user experience. Layout B achieved a higher click-through rate, lower scroll depth, and longer average card dwell time than Layout A. The survey results also indicated that users perceived Layout B as more tailored and visually appealing. However, the application rate did not improve and remained higher for the baseline version. The findings suggest that generative AI can be used to make recruitment interfaces feel more personalized and engaging, but that increased engagement does not necessarily lead to more applications. Due to the relatively small sample size and the use of a prototype with synthetic job postings, the results should be interpreted with caution.