This paper focuses on the optimization of SynergyAI. SynergyAI is a human-AI collaborative system. We addressed challenges related to transparency, efficiency, and advice function in AI programming model. We introduced a visualized decision tree to improve human-AI collaborative capability in terms of explainability. A comprehensive prediction algorithm was proposed to enhance prediction efficiency across data flow. The AI programmer’s advice function utilized scatter-plot matrices for intuitive data relationship visualization. The main contributions of the improved SynergyAI platform are streamlining AI-human collaboration and maximizing prediction accuracy and transparency.