AIConditionalNodeEditor

Overview

The AIConditionalNodeEditor is a component of the Vantage analytics and data platform that allows users to define conditional branches based on AI analysis. It is primarily designed for workflows that require the AI to evaluate input data and direct it to specific branches based on predefined conditions. Each branch can have its own label and description, enhancing the AI's ability to classify data correctly.

Purpose

The primary purpose of the AIConditionalNodeEditor is to facilitate the creation and management of branches, which represent different conditions or outcomes that the AI can classify incoming data into. This feature is critical for users looking to integrate machine learning and AI capabilities into their analytics workflows by providing a structured way to manipulate the flow of information based on AI predictions.

Settings

The AIConditionalNodeEditor has several configuration settings that control its behavior. Below is a comprehensive explanation of each setting:

1. Branches

• Name: branches
• Input Type: Array of objects
• Description: This setting holds all the defined branches. Each branch is an object that consists of a label and an optional description. Users can add, remove, or reorder branches, altering how the AI classifies data. Changing this setting directly influences how many classification options the AI has and how descriptive those options are.
• Default Value: An empty array [].

2. Prompt Template

• Name: promptTemplate
• Input Type: String
• Description: This setting allows users to provide a template for the AI's classification task. The template may include placeholders like {{data}}, {{columns}}, and {{branches}}, which the AI replaces with actual data during processing. Adjusting this setting can refine the AI's focus, potentially leading to better classification results.
• Default Value: An empty string ''.

3. Batch Size

• Name: batchSize
• Input Type: Numeric (integer)
• Description: This setting determines how many classifications the AI processes in a single batch. A higher batch size could improve efficiency but may require more resources and longer processing times. Users need to balance performance and resource allocation.
• Default Value: 10.

4. Classify Column

• Name: classifyColumn
• Input Type: String
• Description: This optional setting specifies a particular column of data that the AI should focus on when making classifications. Providing a specific column can help the AI make more accurate decisions by narrowing its focus. Users can change this setting to test different data inputs.
• Default Value: An empty string ''.

How It Works

1. Branch Management: Users can create, edit, and remove branches. Each branch can be moved up or down in order. Each time a branch is altered, the internal state is updated accordingly.

2. Prompt Template Insertion: Users can insert columns directly into the prompt template, allowing the AI to dynamically incorporate data from specified upstream columns into its classification process.

3. AI Classification: Once branches and the prompt template are configured, the AI analyzes incoming data and uses the provided details to determine the appropriate branch for classification.

Data Expectations

The AIConditionalNodeEditor expects the following types of data:

• Branches: An array of objects, where each object must have at least a label property (string) and can have a description property (string).
• Prompt Template: A string that may contain placeholders related to the workflow.
• Batch Size: An integer defining how many records will be processed in one go.
• Classify Column: A string that denotes a specific column name to provide focused data for the AI.

AI Integrations and Billing Impact

The functioning of the AIConditionalNodeEditor is inherently dependent on the underlying AI services provided by the Vantage platform. As such:

• AI Integrations: The component is designed to connect seamlessly with the platform's AI functionalities, enabling robust classification tasks based on the configuration made in the editor.
• Billing Impact: Usage of this feature might influence billing based on the number of classifications performed, the data size processed, and the complexity of the AI models invoked. Users should monitor their usage statistics to understand potential costs associated with AI operations.

Use Cases & Examples

Use Case 1: Customer Segmentation

A retail company can use the AIConditionalNodeEditor to classify customers based on their purchasing behavior. By defining branches such as "High-value customers", "Occasional buyers", and "New customers", the AI can direct marketing strategies tailored to each segment.

Use Case 2: Order Processing Flow

An e-commerce platform might configure branches representing different order statuses, such as "Pending", "Shipped", and "Cancelled". The AI can then determine the appropriate status based on real-time data and update the branches automatically.

Example Configuration for Customer Segmentation

Objective: Classify customers into different segments based on their purchase amounts.

Configuration Data:

{
  "branches": [
    {
      "label": "High-value customers",
      "description": "Customers who spend over $500"
    },
    {
      "label": "Occasional buyers",
      "description": "Customers who purchase occasionally"
    },
    {
      "label": "New customers",
      "description": "Customers who made their first purchase"
    }
  ],
  "promptTemplate": "Classify this customer's behavior based on their spending history: {{data}}.",
  "batchSize": 15,
  "classifyColumn": "average_purchase"
}

In this setup, the AI utilizes the defined branches and prompt template to analyze each customer and classify them accordingly. Users can observe how different branches are populated based on live data input into the configured workflows.