Mathematics learning in the digital era requires instructional strategies that enhance conceptual understanding while fostering students’ motivation and engagement. Integrating artificial intelligence into classroom practice offers a promising solution to these challenges. This study aimed to examine the effectiveness of integrating an AI chatbot into the Numbered Heads Together (NHT) model based on Problem-Based Learning (PBL) in improving students’ mathematical conceptual understanding and learning motivation. A quantitative approach with a quasi-experimental nonequivalent control group design was employed. The sample consisted of two eighth-grade classes at SMPN 3 Binamu, each comprising 30 students. The experimental group received NHT-based PBL integrated with an AI chatbot, whereas the control group received NHT-based PBL without chatbot support. Data were collected through a conceptual understanding test, a learning motivation questionnaire, and an observation sheet, and were analyzed using independent samples t-test, paired samples t-test, N-Gain, and Cohen’s d. The results revealed a significant difference in learning outcomes between the two groups (p < 0.05). The N-Gain and effect size were moderate, students’ motivation was high, and engagement was very good. These findings indicate that integrating an AI chatbot into NHT-based PBL effectively enhances mathematics learning quality.