Cleft lip and palate (CLP) is a congenital anomaly that can have detrimental effects on an individual's ability to produce certain sounds, including bilabial sounds such as /m/ and alveolar sounds such as /n/, as well as sounds that require closure between the nasal and oral cavities, such as sounds ending in /-ɪŋ/. While surgical intervention can assist in correcting anatomical irregularities, consistent speech and language therapy sessions are often necessary to overcome obstacles associated with speech production. The research problem addressed in this study is the interruption of regular therapy, as evidenced by the discontinuation of services during the COVID-19 pandemic, which can lead to setbacks in progress and an increase in the risk of developmental delays in children. To explore the potential of educational robots to address this problem, this paper (i) investigates the user requirements, usability considerations, and attitudes to speech therapy robots, from the perspective of parents and carers of children with CLP, (ii) presents an experimental prototype for CLP language education and speech therapy using a humanoid robot, named Robot Lily, based on input from prospective users; and (iii) evaluates the prototype Robot Lily on a subject with CLP. The robot has been programmed to interact with the child in a naturalistic manner, using the same language and gestures employed by human speech and language therapists. The robot can respond to inquiries, practice articulation and pronunciation, aid the child with speech fluency, provide feedback, and encourage continued speech exercises. The primary interactions between a child with CLP and Robot Lily were observed and evaluated, with the robot shown to provide a unique opportunity to acquire speech and language abilities in a captivating and interactive manner. Compared to web-based applications, speech education assistant robots like Robot Lily may offer children a more appealing, engaging, and practical solution; and be remotely controlled by humans for specialized interventions at home, thereby circumventing the challenges encountered during COVID-19. These findings and methods have been used to develop an integrated model for ethical robotic speech therapy utilizing machine learning, which is expected to help meet the growing demand for speech therapy services.