The recent developments in the field of large language models (LLMs) have significantly advanced the capabilities of AI in various domains, particularly in customer support, mental health, and proactive dialogues. A notable trend is the integration of LLMs to enhance human-centric applications, focusing on improving the experiences of both service agents and clients. Innovations include systems designed to mitigate emotional stress for customer service agents by transforming the tone of customer messages, and frameworks that construct digital twins of psychological counselors with personalized styles to better meet individual client needs. Additionally, there is a growing emphasis on using LLMs to automate and augment qualitative coding in large unstructured datasets, such as police incident narratives, reducing the need for extensive human intervention while maintaining accuracy. In the realm of proactive dialogues, new frameworks are emerging that automate policy planning from real-world dialogue records, offering more efficient and realistic solutions compared to traditional methods. Furthermore, multimodal systems are being developed to assist in mental health screening, particularly in underserved populations, by leveraging LLMs to interpret complex data like psychological drawings. These advancements collectively push the boundaries of AI's utility and ethical application, aiming to create more personalized, efficient, and accessible services across various sectors.