Agentic FinTech: A Comprehensive Survey on AI Agents in Finance in the Era of LLMs

As large language models (LLMs) evolve from passive text generators into autonomous financial agents with perception, planning, memory, and tool use, a new paradigm-Agentic FinTech-is emerging. This survey formalizes this paradigm by addressing three fundamental questions: what Agentic FinTech is, why it matters, and how such systems can be built and evaluated. We characterize Agentic FinTech through core agentic capabilities, including market-grounded perception, risk-aware planning under constraints, tool use over heterogeneous financial data, long-horizon memory of prior experience, and self-improvement driven by feedback and objectives. From the perspectives of automation and autonomy, we explain why the transition toward agentic systems is essential for next-generation financial intelligence. To support practical implementation and evaluation, we review major financial application domains (e.g., trading, risk management, and regulatory compliance), summarize advances in financial agents in terms of workflows, architectures, and optimizations, and discuss evaluation methodologies, open challenges, and future research directions toward scalable, trustworthy, and efficient agentic financial systems. To keep up with the latest advances in this field, we create a repository at https://github.com/agentic-fintech/Agentic_ FinTech_Survey.