Smartphone AI assistants are gradually becoming an indispensable tool in daily life, and they greatly improve our efficiency and comfort by quickly responding to our commands and needs. It’s essential to enhance user service experience by optimizing voice interaction interfaces and design elements for smartphone AI assistant. The study adopted an experimental method and recruited 140 participants to explore the user experience of visual images (anthropomorphic, abstract, traditional) with different levels of intervention (high, medium, low) of smartphone AI assistant. User satisfaction, likeness, and intention to use were collected through questionnaires. The research results showed that visual image has a significant impact on user preference. Abstract visual image has a higher level of user preference compared to traditional visual image. The level of intervention has a significant impact on user satisfaction, likeness, and intention to use. Smartphone AI assistant with a high level of intervention results in the highest satisfaction, likeness, and intention to use of users. This study explores the effects of visual image and intervention level of smartphone AI assistant on user satisfaction, likeness and behavioral intention, which is of great significance for enhancing the design and human-computer interaction experience of smartphone AI assistant.