This research formulates and tests an AI-integrated e-book to increase English for Specific Purposes (ESP) learning by health science students in order to enhance speaking performance. Conventional ESP materials in health sciences are typically static and non-interactive, limiting opportunities for real-time communicative practice. Using a Research and Development (R&D) design, the project followed four stages: (1) needs analysis involving students, lecturers, and curriculum stakeholders; (2) design and development of AI-integrated e-book; (3) implementation and testing with 58 students through pre- and post-tests; and (4) evaluation and refinement after reflection questionnaires. Fifty-eight health science students completed pre- and post-tests, measured for the fluency, pronunciation, vocabulary, grammar, and content/coherence. The findings revealed a strong demand for AI-integrated ESP materials in need analyses, a notable enhancement in speaking ability (pre-test M = 11.71, post-test M = 19.22; p < 0.001), with improvements noted in all performance areas. Feedback from reflections showed significant satisfaction, emphasizing advantages like accessibility, content relevance, and motivation, while students pointed out issues with internet stability and chatbot constraints. These results validate the benefits of AI integrated e-book into ESP education and provide actionable insights for ESP teachers, curriculum planners, and organizations aiming to integrate interactive AI resources in healthcare education teaching materials. [The page range (118-132) shown on the PDF is incorrect. The correct page range is 118-131.]