The methods consumers use to reduce their perceived risk and make reasonable purchase decisions can be synthesized under the umbrella term “consumer information search behavior” (CISB). As one key factor that conveys a product’s value and quality, price has a significant impact on CISB. There are few studies that comprehensively consider the impact of price level (PL) and perceived price dispersion (PPD) on CISB, and there is a certain disagreement about the impact of PPD specific to the online shopping environment. To address this research gap, we construct a model using the data from 5515 consumers’ purchasing and browsing behavior on a B2C e-commerce website, selecting six products as our research objects. We use a hierarchical regression analysis method to study the influence of product PL and PPD on CISB, and to explore the moderating effect of product categories (durables and consumables) on the relationship between PL, PPD and CISB. The results show that PL significantly affects CISB, and that product categories have a significant moderating effect on the relationship between PL and CISB. For durable goods, when the PL is high, consumers tend to increase their search behavior, both in depth and in breadth, and for consumables with low PL but higher purchase frequency, consumers likewise tend to increase their search behavior. In the B2C online shopping environment, PPD has a significant positive effect on CISB, and product category has a moderating effect on the relationship between PPD and CISB. When consumers purchase consumables, the higher the PPD, the higher the depth of CISB. The findings have several implications for marketing practitioners and enterprises advertising, also can help customers save time and energy in their search behaviors.

Large language models (LLMs) like GPT-4 show potential for scaling motivational interviewing (MI) in addiction care, but require systematic evaluation of therapeutic capabilities. We present a computational framework assessing user-perceived quality (UPQ) through expected and unexpected MI behaviors. Analyzing human therapist and GPT-4 MI sessions via human-AI collaboration, we developed predictive models integrating deep learning and explainable AI to identify 17 MI-consistent (MICO) and MI-inconsistent (MIIN) behavioral metrics. A customized chain-of-thought prompt improved GPT-4's MI performance, reducing inappropriate advice while enhancing reflections and empathy. Although GPT-4 remained marginally inferior to therapists overall, it demonstrated superior advice management capabilities. The model achieved measurable quality improvements through prompt engineering, yet showed limitations in addressing complex emotional nuances. This framework establishes a pathway for optimizing LLM-based therapeutic tools through targeted behavioral metric analysis and human-AI co-evaluation. Findings highlight both the scalability potential and current constraints of LLMs in clinical communication applications.