BrainGrid Controller
Workflow Builder
Events
ChatGPT Event

ChatGPT Event Configuration

The ChatGPT Event enables workflows to leverage OpenAI’s ChatGPT model for automated conversational tasks, data processing, and intelligent responses. This feature is ideal for applications like customer support automation, summarizing text inputs, and interactive responses within a workflow.

Overview

The ChatGPT Event is designed to trigger responses or actions from ChatGPT based on custom prompts. You can configure the event with specific prompts and input mappings, allowing ChatGPT to process and generate responses or computations automatically.

Usage: This event type is used to handle conversational tasks or perform data-based actions, with applications such as:

  • Customer Support: Generate automated responses or summaries to assist with support queries.
  • Text Summarization: Create summaries from large text inputs, such as meeting notes.
  • Automated Replies: Generate automated responses in live chat.

Configuring a ChatGPT Event

Step 1: Set Up the OpenAI Token

  1. OpenAI Token: Input your OpenAI API token to authenticate the ChatGPT event and enable requests to the OpenAI API.

Step 2: Create the Prompt

The prompt can carry data from previous blocks using "@" to reference specific fields or data points. Customize the prompt to specify the kind of output you expect ChatGPT to generate.

Example Prompt

Below is an example prompt for a ChatGPT event designed to evaluate a student's answers against rubric criteria and generate a detailed response in JSON format:

Looking into the rubrics criteria: 
@create exam record Rubrics Criteria 
and evaluation matrix:
@create exam record Evaluation Matrix 
I want you to review the answer sheet of the student
@create exam record Student Answer 
We need more detailed evaluation and comments. Need to strictly answer all questions.
Give output as JSON for the marks obtained by the student, including the total marks of the examination and the credit category they fall under.
{
  "answers": [
    {
      "question": "(question_name)",
      "criteria": "",
      "student_answer": "",
      "marks_obtained": "(float)",
      "total_marks": "(float)",
      "comments": "Good comparison, but lacks critical depth."
    },
    {
      "question": "(question_name)",
      "criteria": "",
      "student_answer": "",
      "marks_obtained": "(float)",
      "total_marks": "(float)",
      "comments": "Good comparison, but lacks critical depth."
    }
  ],
  "total_marks_obtained": "(float)",
  "total_max_marks": "(float)",
  "percentage": "(float)",
  "credit_category": ""
}

Step 3: Validate and Test

  • Action: Run the ChatGPT event with test data to ensure the output matches the expected format.
  • Verification: Check that ChatGPT processes inputs correctly and generates responses that align with the specified prompt.

Step 4: Deploy the Event

  • Deploy: Once validated, deploy the ChatGPT event within the workflow.
  • Review Output: Confirm the output is generated as expected, and adjust the prompt if necessary for accuracy.

Summary

The ChatGPT Event allows workflows to incorporate AI-driven responses, enriching automation capabilities with advanced conversational or processing logic. By configuring a prompt tailored to your specific needs, you can automate tasks like evaluating data or providing real-time assistance in workflows.

Webhook EventExternal Database Event