3 Working with generative AI in finance
This chapter covers
- The shift from model building to AI engineering
- Secure, compliant GenAI systems for finance
- Model selections beyond generic benchmarks
- RAG, PEFT, and guarded AI agents techniques
- Strategies to manage cost, risk, and governance
Finance has always gravitated toward data-driven innovation, yet the debut of generative AI (GenAI) marks a decisive turning point. Until recently, most machine learning (ML) in banks and asset management firms focused on structured data analytics, such as credit scoring. Then came the proliferation of large language models (LLMs), tools capable of generating human-like text, writing code, or summarizing dense regulatory documents in seconds. This leap has triggered a gold rush in financial services, with institutions like Morgan Stanley embedding chat-driven experiences into their advisory platforms.
The influence goes far beyond automating back-office forms. These conversation-driven, generative tools inspire both excitement and caution, demanding a new AI lens that considers not only the technology but also the guardrails needed to align it with business strategy and regulation.