Skip to content

Getting Started

In this guide, we'll demonstrate how to safeguard and monitor conversations in LLM-powered applications using Trust3.

As an example, we will build a simple Intelligent Sales Assistant using Snowflake Cortex Agents and Streamlit. This assistant simulates realistic LLM interactions and helps showcase how Trust3 can:

  • Enforce AI guardrails
  • Monitor usage
  • Ensure secure and responsible AI behavior

Let’s get started!

Github

All the files mentioned in this quickstart guide — whether to download or copy — are available in the following GitHub repository:
Trust3 Snowflake Client Example

What You’ll Learn

  • How to install and configure the AI Trust Layer for Cortex from the Snowflake Marketplace

  • How to integrate Trust3 with your Streamlit app to:

  • Enforce content and access guardrails

  • Monitor LLM interactions
  • Ensure responsible and secure AI usage

What You’ll Build

  • A Sales Assistant Application that allows users to:

  • Interact with sales data via semantic search

  • Perform metrics analysis
  • Ask questions using LLM-powered Q&A

  • End-to-end safeguards and monitoring of all assistant conversations using Trust3 to ensure:

  • Security

  • Compliance
  • Responsible AI behavior

What You’ll Need

Snowflake Account

  • Permissions to:

  • Create roles

  • Create databases and tables
  • Upload files
  • Create external access integrations
  • Install applications from the Snowflake Marketplace

Access to Snowflake Cortex Services

  • Cortex Agents
  • Cortex Search
  • Cortex Analyst

AI Trust Layer for Cortex

  • Available via the Snowflake Marketplace
  • Required to safeguard and monitor LLM assistant conversations through Trust3

What Next?