Today's digital context brings a diverse matrix of challenges to optimize CSP operations while enhancing customer experience. High call volumes, complex architectures, diverse layers of technical issues, and insufficient-inefficient tools overwhelm human agents, leading to decreased productivity, and low quality of life for customer-facing workforce. To address these challenges, the Catalyst project harnesses the power of AI to provide CSPs with practical solutions that intelligently extract poignant and actionable insights from conversational data. It empowers virtual agents to make intelligent decisions using real-time data. This Catalyst will explore and jointly innovate: 1. Data-driven self-learning capability: Explore high-quality customer service experience from historical data, including reply words and tool operation. Compared with mainstream operation management, it mainly relies on manual experience summary and robot process configuration and operation. It can solve the problem of high manual investment and low efficiency. In addition, big data accumulation and learning experience can effectively ensure the consistency of customer service experience. 2. Experience learning enables intelligent virtual customer service: High-quality experience mined from data can not only assist and guide agent operations, but also directly enable virtual customer service customers. By accurately understanding customers' intentions, accurately answering customers' questions, and flexibly invoking a third-party system, the solution helps customers quickly resolve problems, reduces the workload of manual agents, reduces costs, improves efficiency, and improves customer service experience. Virtual assistants & co-pilots are being tested and considered across the industry with some already well-established use cases. However, with the growth of GenAI, it is important to continue to drive these discussions. This project would provide a valuable use case to show how these AI virtual assistants can innovate the telecoms industry.