PredictiveInsightsTile Documentation

Overview

The PredictiveInsightsTile is an AI-powered visualization component designed to provide predictive insights based on workflow data and user-defined prompts. It generates predictions, analytics, and actionable recommendations, caching results for efficient access within a user session. This component integrates with Vantage's dashboard capabilities, making it a powerful asset for data-driven decision-making.

Purpose

The primary purpose of the PredictiveInsightsTile is to leverage AI models to process workflow data and derive actionable predictions that help users anticipate outcomes and make informed decisions. The tile supports a variety of use cases in business analytics, operational efficiency, and strategic planning.

Settings

The configuration for the PredictiveInsightsTile includes several essential settings:

Title

• Name: title
• Input Type: String
• Description: Indicates the title displayed at the top of the tile. This setting influences how users recognize the purpose of the tile.
• Default Value: 'Predictive Insights'

Subtitle

• Name: subtitle
• Input Type: String
• Description: A supplementary description or label that provides additional context about the predictions the tile will generate. This helps in better understanding the displayed insights.
• Default Value: undefined

Footer Text

• Name: footerText
• Input Type: String
• Description: Text displayed at the bottom of the tile. This could include disclaimers or additional information about the AI predictions.
• Default Value: undefined

Footer Color

• Name: footerColor
• Input Type: String (CSS color format)
• Description: Defines the background color of the footer. Changing this value will alter how the footer visually integrates with the overall design of the tile.
• Default Value: undefined

Workflow ID

• Name: workflowId
• Input Type: String
• Description: Unique identifier for the workflow the predictions will be generated against. This parameter is crucial as it determines which data context will be utilized for prediction generation.
• Default Value: undefined

Prediction Prompt

• Name: predictionPrompt
• Input Type: String
• Description: A user-defined prompt guiding the AI in generating predictions. This is essential for directing the AI model on the nature of insights to be produced.
• Default Value: undefined

Source Data Snapshot

• Name: sourceDataSnapshot
• Input Type: Array of objects
• Description: Contains pre-captured data relevant to the tile, stored when the tile is created. This data serves as a fallback if real-time data is unavailable, providing context for the AI's prediction generation.
• Default Value: undefined

How It Works

The PredictiveInsightsTile operates by generating an AI prediction based on the provided workflow data and prediction prompt. The component performs the following steps:

1. Initialization: On mounting, the component checks if a valid prediction prompt is provided.
2. Data Caching: It leverages session-level caching to store predictions across tab switches, avoiding unnecessary API calls for previously generated predictions.
3. Generate Prediction: If no cached data exists, it fetches real-time predictions using the configured workflow ID and source data snapshot by making an API call to /api/ai/predictive-insights.
4. Display Results: The component updates its state with the response data, displaying predictions, key analytics, and actionable insights. If errors occur, it presents an error message to the user.

Data Expectations

The PredictiveInsightsTile expects the following data:

• The workflowId specifies which data context to use for generating predictions.
• The predictionPrompt is a string that directs the AI on what type of insights to produce.
• sourceDataSnapshot is an optional array containing historical or sample data relevant to the predictions.

AI Integrations

This tile is integrated with an AI-driven backend, specifically fetching predictions based on data and user prompts. Depending on the configuration, it can utilize different AI models (e.g., OpenAI or Claude). The data returned may vary based on the model used, and the component dynamically displays this information in the footer as an acknowledgment of the AI provider and model used.

Billing Impact

Utilizing the PredictiveInsightsTile may impact billing depending on the model serving the prediction. Each API call made to generate predictions may incur costs associated with AI processing and data retrieval. Users should consult Vantage’s billing policies to understand how this component may factor into their usage pricing.

Use Cases & Examples

Use Case 1: Sales Forecasting

A sales department might use the PredictiveInsightsTile to generate predictions about future sales based on historical sales data and seasonality trends. By inputting a prompt such as "Forecast sales for the next quarter," the tile can assist in strategic planning.

Use Case 2: Operational Efficiency

An operations manager can leverage the tile to predict potential bottlenecks in production by analyzing time-series data of different production stages. An appropriate prompt could be "Identify possible delays in the production line."

Use Case 3: Customer Segmentation

A marketing team can utilize the PredictiveInsightsTile to analyze customer interactions and behaviors, generating predictions about customer segments most likely to respond to a new campaign.

Example Configuration

Scenario: Sales Forecasting

In this example, the PredictiveInsightsTile is configured to predict sales for the upcoming quarter.

{
  "title": "Quarterly Sales Predictions",
  "subtitle": "AI-generated sales forecast for Q2",
  "workflowId": "sales_forecast_2023",
  "predictionPrompt": "Forecast sales for the next quarter based on historical data.",
  "sourceDataSnapshot": [
    { "month": "January", "sales": 15000 },
    { "month": "February", "sales": 18000 },
    { "month": "March", "sales": 20000 }
  ],
  "footerText": "Predictions are generated based on historical data trends.",
  "footerColor": "#f0f9ff"
}

This configuration allows the sales department to visualize predictions effectively, enhancing their decision-making capabilities for the upcoming quarter.