Workflow to Build an AI Agent

Here's a summary of the steps we'll follow to build the AI agent.

Step Tasks
Complete Prerequisites

Get connection information for the Large Language Model (LLM) you'll be using. You'll need this information to configure your LLM for your AI agent.

Required information:

  • Base URL for the LLM.

    Example: https://api.openai.com

  • Model to use.

    Example: gpt-40-mini

  • API Key for the LLM
1

Plan Your AI Agent. Take a few moments to plan your AI agent. This will help you to clearly define the AI agent's purpose and which tools it will need.

2

Create the Project for the AI Agent.

You create AI agents inside projects.

3

Create integrations optimized for AI agents.

In order to use an integration as an agentic AI tool, the integration must have a REST trigger, a POST verb, and JSON input and output.

  1. Create the integration, activate it, and register it as an agentic AI tool.
    1. Create the Integration: Auto Approve Expense Report
    2. Register the Integration as an Agentic AI Tool

4

 Build the AI agent.
  1. Configure the Agentic AI Thinking Pattern. Define which thinking pattern your AI agent will use and configure your LLM connection information.
  2. Add and Configure the AI Agent.
5 Run and Test the AI Agent.

Next step: Plan Your AI Agent.