What are the challenges of integrating AI in finance and investment research?

AI agents are becoming increasingly popular in finance, but they face significant challenges. One major issue is the quality of data; much of the data available is messy and unstandardized, making it difficult for AI to perform effective analysis. Additionally, AI agents can become bogged down in data cleaning rather than focusing on hypothesis validation. How do you think we can overcome these challenges? Have you encountered similar issues in your projects?