Many companies possess valuable internal data, such as customer interaction analytics, audit logs, and support tickets, which can highlight areas needing attention and feature requests. Previously, extracting insights from such data required specialized knowledge and custom model training, but large language models (LLMs) now allow for simpler processes using prompts. The text outlines how to build an internal tool using Streamlit to enable employees to experiment with LLMs on datasets. This involves logging in, writing prompts for ticket classification, testing prompts, and saving them for others. Streamlit facilitates the creation of interactive data applications by providing components for data visualization and interaction, as well as utilities for data loading, managing secrets, and caching. Authentication is managed through PropelAuth, which integrates with Streamlit to provide various login options and protect data access. The tool allows users to save prompts using existing database connections, streamlining the development of AI-driven data applications that are secure and tailored to organizational needs.