This survey paper provides a comprehensive overview of the evolution of artificial intelligence in quantitative investment, with a particular focus on alpha strategy as a representative example. It traces the progression from early-stage quantitative research, which relied on human-crafted features and traditional statistical models within an established alpha pipeline, to the transformative impact of deep learning. Deep learning techniques have enabled scalable and integrated modeling across the entire quantitative investment pipeline, from data processing to order execution, enhancing predictive capabilities and efficiency.

Building on this foundation, the paper highlights the emerging role of large language models (LLMs) in pushing AI beyond mere prediction. LLMs empower autonomous agents to process unstructured data, generate alphas, and support self-iterative workflows, suggesting a potential paradigm shift towards agent-based automation in quantitative finance. This survey underscores how these AI advancements are reshaping the field, moving from human-driven processes to more automated, intelligent systems that could redefine quantitative investment strategies and operations.