Security in Enterprise AI
Learn how to create a powerful financial research agent with custom prompts, user inputs, and dynamic outputs—all in one workflow.
In this hands-on session, you’ll build a financial analyst AI agent that processes SEC filings like 10-Ks and 10-Qs. You’ll learn how to start from a blank workflow, configure LLMs, define custom prompts, and design a user interface with file uploads, verbosity controls, and input instructions. By the end, your agent will be able to extract insights, cite references, search a knowledge base, and export results as PDF, Word, or Excel—automatically.
Summary
Create a StackAI workflow from scratch using the workflow builder
Upload a 10-Q or 10-K financial file via a file input node
Configure a large language model (e.g. Gemini) and connect it to the file input
Write a custom prompt for financial analysis with references and citations
Add user inputs for custom instructions and verbosity level (e.g. bullet points, deep analysis)
Build a structured form-based UI with inputs and dynamic outputs
Add a second LLM agent focused on Management Discussion & Analysis
Connect both analysts to a SharePoint knowledge base of past reports
Run and test the agent with live inputs and see model output in real time
Enable multiple outputs for different analytical focuses
Customize the UI with branding, SSO, and access controls
Export results in PDF, Word, Excel, or plain text formats
View usage in the Analytics tab and monitor real-time activity in the Manager view
Next step: Learn to use the Quick Start flow to build a GPT-style assistant in the following video