AI Hub

AI Hub is designed to help you easily discover insights in your comment data collected from sources outside of the Perceptyx platform (e.g., Glassdoor, LinkedIn, customer feedback). It uses four proprietary AI models to automatically analyze your data for you in a simple report you can use to quickly identify what’s going well and where you might need to focus action. This tool is specifically tailored for data scientists and people analytics teams to enhance comment analysis capabilities, enabling deeper insights and data-driven decision-making.

   Note: AI Hub is available only for Administrator users.

This article walks through:

• Access AI Hub
• Navigate AI Hub
• Upload Data for Analysis
• Manage Your Files
• AI Hub Models
• Report Format and Output

Access AI Hub

1. Log in to the People Insights Platform. 

2. In the navigation bar, click Analyze.

The Analyze page appears.

3. Click AI Hub (upper right).

The AI Hub landing page appears.

Once you arrive at the AI Hub, you can upload your datasets and harness the power of artificial intelligence to process and analyze it. Our advanced AI algorithms process and interpret the data, revealing meaningful patterns and trends.

Navigate AI Hub

The AI Hub is organized into four main sections: side navigation panel, Overview, Upload Your Data, and File Manager. Each area is described below.

Side Navigation Panel

Navigate easily around AI Hub using the side navigation panel, which is expanded by default (you can click the arrow to collapse it). This panel is always displayed in AI Hub and enables you to click the icons to access the AI Hub menu options: Overview, Upload Data, File Manager, and Help.

Overview (AI Hub Main Area)

The Overview area features access to learn more about the different model types. It also enables you to access the Upload Data and File Manager areas by clicking the associated button.

Upload Your Data

You can upload your data for analysis in the Upload Data page, which provides instructions for preparing your data file and a link to download a template you can use as a guide. Using the Select File feature, you select your prepared file, review your selection, and upload the data. Once the data is uploaded, the file is processed  to interpret the data, revealing patterns and trends. A progress indicator at the top of the page updates as you work through the upload process. 

Note: Your file may take up to 24 hours or more to process depending on size and content.

File Manager

You can review, download, rename, and delete your uploaded files in File Manager. Your files list with status, uploaded date, and action buttons. You can sort your files by file name, status, and upload date by clicking the up or down arrows. You can select an action for a single file, or select multiple files and apply an action.

Help

Clicking the Help option opens the Perceptyx Help Desk portal in a separate browser window. You can use the portal to submit and track support tickets and to access Knowledge Base articles (how-to instructions and best practices).

Upload Data for Analysis

1. Navigate to the AI Hub.

2. Click Upload Your Data in the Overview area (or Upload Data in the side navigation panel).

The Upload Your External Data page appears with instructions for preparing your data file.

You have the option to upload your data file or download a template and populate it with your data. If this is your first time uploading, we recommend you download the template and populate it to ensure the format is correct.

Note: The data file must be in .csv format and include two columns of information: Comment ID (number 0-X) and Text (comment text).

3. Download the template (if needed).

   1. Click Download CSV Template.

   2. Locate and open the downloaded template, populate it with your data, and save it in .csv format.

4. Upload your data.

   1. Click Select File to Upload.

   2. Navigate to your data file and double-click the file name.

The selected file displays in the Upload Your External Data page.

3. Click Upload and Process.

If the upload is successful, you will receive an Upload Successful message.

If there is an issue with your file, a detailed error message will display with a suggestion about how to correct the issue. For example:

4. If you receive an error message, follow the instructions and upload your file again.

You will receive an email when the file has been processed. This can take up to 24 hours (or more), depending on the file size and content.

Manage Your Files

After you have uploaded a file, you can access, download, rename, or delete the file(s) using File Manager.

1. Navigate to the AI Hub.

2. Click File Manager (left panel).

The File Manager page displays.

3. Identify the file you want to download, rename, or delete.

4. Click the file’s Actions dropdown and select the desired action:

   • Download

     1. On selection, the system downloads your file. 

     2. Navigate to your Downloads folder and open the file.

   • Rename File

     1. On selection, the Enter New Name field displays. 

     2. Enter the new name and click Submit.

   • Delete
     On selection, the system deletes your file.

AI Hub Models

AI Hub uses four models to process and interpret data:

• Emotion: Measures nuanced emotions expressed in comments, including anger, fear, love, joy, sadness, and surprise. 

• Themes: Organizes comment data by topics or themes for efficient categorization and retrieval.

• Intent: Extracts the intent behind each comment, to further categorize text not only by topic, but also purpose. 

• Sentiment: Gauges the positivity of a piece of text, on a calibrated sliding scale from low sentiment to high sentiment.

You can learn more about each model by clicking Learn more.

Report Format and Output

The processed data file is delivered in .csv format with separate columns for each theme, intent, and emotion, as well as a Sentiment column.

• For every comment in your dataset, the processed file shows a 0 or a 1 to indicate whether a specific theme, intent, or emotion was detected.

• The Sentiment column displays Positive, Neutral, or Negative to indicate the type of sentiment detected.

• You can easily aggregate the data for any theme, intent, or emotion.