Creating an AI Agent for Financial Report Analysis

Resources (including link to code along notebook): https://bit.ly/41cgavS

AI agents are transforming industries by automating complex processes and delivering insights at scale. In financial services, AI agents can streamline decision-making, reduce manual effort, and improve the accuracy of report analysis can informs various downstream tasks like market research, credit scoring, report generation, etc. Designing and building such agents requires a strong understanding of their architecture, the data they rely on, and how to use AI to automate repetitive tasks effectively.

In this hands-on code-along session, Jayeeta Putatunda, a Lead Data Scientist & Director at Fitch Group, guides you through creating an AI agent tailored for financial report analysis. You’ll learn how to design and architect AI agents, explore their applications in finance, and identify the key data needed for these systems. The session will also cover how AI agents can automate repetitive tasks to enhance efficiency. This webinar is ideal for data scientists and machine learning scientists looking to build practical AI solutions for financial applications.

00:00 Introduction & Welcome
00:21 Why AI Agents for Financial Reporting?
01:44 Guest Introduction – Jayta from Fitch Group
03:27 Understanding AI Agents vs. Agentic AI
05:56 Identifying Valuable Use Cases for AI Agents
07:44 Key Components of an AI Agent
10:58 Choosing the Right AI Agent Approach
12:19 AI in Financial Services – Real-World Applications
13:55 Today's Use Case: Financial Report Analysis
16:05 Setting Up the AI Agent Workflow
18:34 Required Tools & API Setup (Grok & Agonal)
22:06 Agent 1: Web Search-Based Research Agent
26:14 Running the Research Agent – Example Queries
31:51 Agent 2: Retrieval-Augmented Generation (RAG)
35:16 Setting Up Vector Database for RAG
38:53 Loading & Processing Financial Documents
42:30 Running Queries Against the Knowledge Base
44:27 Agent 3: AI-Driven Stock Market Analysis
47:40 Running Market Comparison & Trend Analysis
50:56 Agent 4: Automated Evaluation Framework
54:40 Reviewing Evaluation Metrics & Results
57:10 Best Practices for AI Agent Development
59:50 Q&A – Choosing the Right Vector Database
01:02:26 Q&A – LangChain vs. Agonal for AI Agents
01:05:02 Q&A – How AI Agents Improve Financial Workflows
01:10:28 Closing Thoughts & Upcoming Sessions