Skip to main content

Train Neo

Training Neo is a powerful feature that allows you to refine and improve the AI's understanding and performance for your specific workflows. By providing feedback on past workflow executions, Neo can automatically suggest and apply improvements to the custom instructions of the actions within those workflows. This leads to more accurate and tailored automations over time.

Why Train Neo?​

  • Enhance AI Accuracy: Teach Neo how to better interpret situations and data specific to your environment.
  • Tailor Workflow Behavior: Guide Neo to make decisions and take actions that align more closely with your business processes and preferences.
  • Improve Automation Outcomes: Get more precise and effective results from your automated workflows by refining the AI's instructions.
  • Iterative Improvement: Continuously make Neo smarter and more helpful through an interactive feedback loop.
tip

The "Train Neo" feature primarily focuses on enhancing the custom instructions associated with actions in your workflows. While Neo can automatically suggest these improvements, you can also manually update custom instructions directly within each action's configuration settings in the workflow editor.

How It Works​

The training process involves a few key steps:

  1. Identify an Execution: You select a past workflow execution where the outcome could be improved.
  2. Provide Feedback: You give specific feedback on what went wrong or how the AI could have performed better for that particular execution.
  3. Neo Analyzes: Neo's Workflow Improver agent analyzes your feedback, the context of the ticket (if applicable), the workflow's last execution message, and the current custom instructions of the involved actions.
  4. Neo Suggests Improvements: Based on its analysis, Neo proposes specific changes to the custom instructions for the relevant actions. It will also provide a reasoning for its suggestions.
  5. Review and Apply: You review the suggested improvements. You can choose to accept and apply them directly. Applying the changes updates the custom instructions within the workflow's actions.

Step-by-Step Guide to Training Neo​

Let's walk through an example of how to train Neo using feedback from a "Triage" workflow.

1. Navigate to Workflow Executions​

From the NeoAgent dashboard, go to Workflows. You'll see a list of your "Triggered Workflows". Identify the workflow you want to improve (e.g., "Triage"). Click on the "Event History" icon for that workflow to view its execution history.

Accessing Workflow Event History from the Workflows List

2. Select a Specific Event Execution​

You'll be taken to the "Event Executions" page for that workflow (e.g., "Event Executions - Workflow 4941"). Here, you can filter by Ticket ID or Ticket Number to find the specific execution you want to provide feedback on. Once you've found the relevant execution (e.g., for Ticket ID 1240), click on the "Train Neo" icon (lightbulb icon) in the "Actions" column for that run.

Selecting a specific event execution and clicking the Train Neo icon

3. Provide Your Feedback​

A modal dialog titled "Train Neo to get better" will appear.

  • It will confirm which actions within the workflow have custom instructions capability and can therefore be improved (e.g., "Workflow can be improved. The following actions have custom instructions capability: Ticket Triage").
  • You'll see the status of the selected execution (e.g., "Execution Status: SUCCESS").
  • In the "Your Feedback" text area, provide specific details about what Neo should have done differently or what it missed. The more context you provide, the better Neo can learn.

Example Feedback: "This ticket should have 'Printer' type because it's related to the HP device in room 2."

Providing feedback in the Train Neo modal

Click SUBMIT FEEDBACK.

4. Review Suggested Improvements​

After you submit your feedback, Neo will analyze it and present you with "Review Suggested Improvements." This dialog shows:

  • Your Feedback: The feedback you just provided.
  • Neo's Analysis: An explanation of how Neo interpreted your feedback and what it plans to do. Example Analysis: "Human feedback indicates that the ticket should be categorized as 'Printer' type due to references to an HP device in room 2. The 'type' field previously did not include any custom instructions. The improved instructions now explicitly check for printer-related keywords and assign the 'Printer' type accordingly, while preserving the existing default behavior for other scenarios."
  • Suggested Improvements: A list of specific changes Neo proposes for the custom instructions of relevant actions and their fields. Example Suggestion for "Ticket Triage" action, "type" field: 
    + New instructions added
    - Review the ticket content for printer-related keywords such as 'HP', 'Printer', or 'room 2'.
    - If any of these indicators are present, assign the ticket type as 'Printer'.
    - If no printer-specific indicators are detected, maintain the default action of leaving the field as None.

Reviewing Neo's suggested improvements to custom instructions

Carefully review Neo's analysis and the suggested changes to the custom instructions.

5. Apply Selected Improvements​

If you are satisfied with the suggested improvements, click the APPLY SELECTED IMPROVEMENTS button. Neo will then update the custom instructions for the specified actions within your workflow. The workflow will use these updated instructions for future executions.

Important Considerations​

  • Focus on Custom Instructions: The "Train Neo" feature is designed to refine the custom_instructions field within your workflow actions. This is where you provide plain English guidance to the AI.
  • Manual Updates: Remember, you can always manually edit the custom instructions for any action directly in the workflow editor if you prefer, or if you want to make changes outside of the feedback loop.
  • Context is Key: Provide clear, specific, and contextual feedback. The better your feedback, the more effectively Neo can learn and improve.
  • Iterative Process: Training Neo is an ongoing process. The more feedback you provide over time, the more aligned Neo will become with your specific operational needs.
  • Check Improvability: Not all actions or workflows may be improvable through this feature. An action must have custom instruction capabilities. The system will inform you if a selected workflow/action can be improved.
tip

The more you interact with and guide Neo, the smarter and more helpful it becomes.