/
Creating an AI Agent

Creating an AI Agent

  • Verified

    • This page covers creating a new AI Agent in the Profound AI IDE. For information on how to deploy the Agent, see here.

    Overview

    Creating an AI Agent is the first step in embedding conversational or logic-driven intelligence into your Profound AI applications.

    Through the IDE interface, you can define a new Agent’s parameters, preview its behavior, and validate its conversational logic before deployment. This process helps ensure that Agents behave as expected and align with your application goals and user interactions.

    This page includes both a video walkthrough of creating an AI Agent and step-by-step written instructions.

    Video Walkthrough

    Steps

    1. Open the IDE using the URL formatted as follows: http://systemName:port/ide.

      1. The systemName and port mentioned in the URL above will be specific to your system and instance port and should be replaced with the correct values.

        • systemName - Should be replaced with the correct system name (or IP address) for your system.

        • port - Should be replaced with the port that your instance was installed on.

      2. Make sure your Profound AI instance is running and accessible.

    2. In the IDE menu, select New → AI Agent. This opens the Agent setup dialog.

       

    3. Next, provide the appropriate Agent parameters, for example:

       

    4. An Agent Preview will appear in the center of the IDE showing a simulated chat window. This allows you to see how the Agent might respond to inquiries.

       

    5. Test the AI Agent’s functionality by interacting with the Agent Preview.

      1. Enter test messages to the Agent. Observe the responses, latency, and conversational flow.

        1. Watch for edge cases, misunderstood inputs, or undesired outputs.

      2. Retest in the Preview until the behavior is acceptable.