Plan Your AI Agent

Before we start creating the AI agent, let's identify the agent purpose and which tools the AI agent will use. This will help you have all the required information when creating the AI agent and tools in Oracle Integration.

What's the Purpose of the AI agent?

For this tutorial we're creating a simple AI agent to approve or reject expense claims.

AI Agent	Purpose
Process Expense Claims	Approve or reject expense claims. Automatically approve expenses equal to or less than $50 USD. Reject the expense if the expense is more than $50 USD.

What Tools does the AI Agent Need?

The AI agent needs one integration to automatically approve expense claims. Since all the reasoning is done in the AI agent, the integration can be very simple.

Identify which tools are needed, their purpose, input and output parameters.

Here's an overview of the tools we need for the AI agent:

Tool Name Purpose Input Parameters Output Parameters

Auto Approve Expense Report

Automatically approve submitted expenses.

Tool Name	Purpose	Input Parameters	Output Parameters
Auto Approve Expense Report	Automatically approve submitted expenses.	`AmountClaimed`: string. This is the dollar amount submitted for reimbursement. Example JSON: `{ "AmountClaimed" : "" }`	`AutoApproveResponse`: string. Indicates whether the expense has been approved. Value is always Approved since this integration is only used for automatic approval. Example JSON: `{ "AutoApproveResponse" : "" }`

AmountClaimed: string. This is the dollar amount submitted for reimbursement.

Example JSON:

{ "AmountClaimed" : "" }

AutoApproveResponse: string.

Indicates whether the expense has been approved. Value is always Approved since this integration is only used for automatic approval.

Example JSON:

{
  "AutoApproveResponse" : ""
}

Next step: Create the Project for the AI Agent. Now that we know what the agent does and the tools that we need, we're ready to start building the components.