This study proposes a method for laptop recommendation in a conversational recommender system (CRS) by integrating collaborative filtering with the Apriori algorithm. The CRS interacts with users to help them find laptops that match their preferences, allowing them to provide feedback or critiques on the recommendations. This research emphasizes the use of compound critiques, which allow users to express preferences on multiple attributes at once, leading to more personalized recommendations. The Apriori algorithm identifies frequent itemsets from these critiques, which are then used to iteratively update recommendations. Evaluation results show that the High Support (HS) strategy, which focuses on commonly preferred features, produces more efficient recommendations, with a shorter average session duration of 38.01 seconds compared to the Low Support (LS) 41.30 seconds and Random (RAND) 50.49 seconds. This approach improves the recommendation process by better aligning with user preferences, which in turn improves interaction efficiency.