探討個人化、訊息類型及論述品質對於AI聊天機器人之消費者行為影響

Abstract

隨著智慧型移動裝置普及，AI聊天機器人的功能興起，消費者能透過觀察店家評論，並選擇評價良好之店家做為消費前之重要考量。因此，本研究採2（個人化程度：高個人化vs. 低個人化）x2（訊息類型：單面訊息vs. 雙面訊息）x2（論述品質：強論述vs. 弱論述）三因子組間實驗設計，探討三者在線上評論中對於消費者行為之影響，以及資訊可信度感知在自變項（個人化程度、訊息類型、論述品質）與依變項（產品態度、購買意圖、口碑推薦意願）之間的中介效果。研究方法為實驗法，透過網路問卷讓281位參與者被隨機分配到不同的實驗，並被要求觀看相應的刺激物。本研究結果證實，（1）高個人化聊天機器人比低個人化聊天機器人，引發消費者較佳的產品態度、購買意圖及口碑推薦意願；（2）單面訊息評論比雙面訊息評論，引發消費者較佳的產品態度、購買意圖及口碑推薦意願；（3）強論述評論比弱論述評論，引發消費者較佳的產品態度、購買意圖及口碑推薦意願；（4）對瀏覽單面訊息之消費者而言，搭配強論述比搭配弱論述，產生較佳之產品態度、購買意圖及口碑推薦意願；（5）使用高個人化聊天機器人，搭配單面訊息並強論述評論，產生較佳之產品態度、購買意圖及口碑推薦意願；（6）資訊可信度感知在個人化、訊息類型、論述品質與產品態度、購買意圖及口碑推薦意願之間存在中介效果。

With the popularity of smart mobile devices, AI chatbots have emerged as a key tool for consumers seeking information. Consumers are accustomed to observing store reviews and choosing stores with good reviews as an important consideration before purchasing. This study employed a three-factor between-subjects experimental design two (personalization level: high vs. low) x two (message type: one-sided vs. two-sided) x 2 (argument quality: strong vs. weak) to examine the influence of these factors on consumer behavior in online reviews. Furthermore, it investigated the mediating role of perceptions of information credibility between the independent variables (personalization level, message type, and argument quality) and the dependent variables (product attitude, purchase intention, and word-of-mouth recommendation).The study adopted an experimental methodology, in which 281 participants were randomly assigned to different experimental conditions via an online questionnaire and instructed to view corresponding stimuli.The results showed that: (1) Chatbots with high personalization generate better product attitude, purchase intention, and word-of-mouth recommendation compared to chatbots with low personalization; (2) One-sided reviews generate better product attitude, purchase intention, and word-of-mouth recommendation than two-sided reviews; (3) Reviews with strong argument lead to better product attitude, purchase intention, and word-of-mouth recommendation compared to those with weak argument; (4) Among consumers exposed to one-sided messages, reviews with strong argument generate better product attitude, purchase intention, and word-of-mouth recommendation than reviews with weak argument; (5) Highly personalized chatbots delivering one-sided messages with strong argument generate better product attitude, purchase intention, and word-of-mouth recommendation; (6) Perceptions of information credibility mediates the effects of personalization, message type, and argument quality on product attitude, purchase intention, and word-of-mouth recommendation.