Contact Us

If you still have questions or prefer to get help directly from an agent, please submit a request.
We’ll get back to you as soon as possible.

Please fill out the contact form below and we will reply as soon as possible.

    English (US)
    MX Spanish (Mexico)
    CA French (Canada)
    US English (US)
    • Home
    • Employee Communications
    • Communications App

    Generated summaries

    Written by Lindsey Stanifer

    Updated at November 4th, 2026

    Contact Us

    If you still have questions or prefer to get help directly from an agent, please submit a request.
    We’ll get back to you as soon as possible.

    Please fill out the contact form below and we will reply as soon as possible.

    • Getting Started
      Setting up your calendar Managing your alerts Additional Settings
    • Daily Processes
      Candidate Inbox Candidate Profile My Calendar/Calendars Browser Extension My Jobs Approvals Engine Manual and Common Scheduling Practices Form I-9 processes Forms & Offers Recruiting Dashboard Paradox Assist
    • Candidate and User Engagement
      Campaigns Conversation Builder Surveys Channels/Phone Numbers Talent Community Voice Web Management Analytics & Reporting Data Privacy Career Sites Multilingual Source Tracking
    • Management Tools
      Job Management Scheduling Security Journeys Content Management System (CMS) Multiple Brands Workflows Users, Roles and Permissions Data Feeds Location Management Lookup Tables Assistant Messaging Company Information Client Setup System Attributes Integration Center Workday Implementation
    • Contextual AI
      Knowledge Training Library
    • Conversational Events and Campus
      Conversational Events Campus Events
    • Employee Communications
      Communications App Employee Management
    • Release Notes
      2026 2025
    • Workday Feature Descriptions
    • Troubleshooting
    + More

    Table of Contents

    Configuration Less than 50 responses experience 50+ responses experience Summary refresh Keywords Keyword generation logic Exporting responses Special case: Single-select questions Employee surveys User feedback Permissions

    This feature is used in conjunction with the Employee Communications product. If you have not purchased this product, then you will be unable to access this feature and its functionality.

     

    You do not need Contextual AI enabled in order to use this feature.

     

    Generated summaries are AI-generated descriptions of employee responses when at least 50 people have responded to an active communication. This assists users in gathering a general understanding of employee feedback while the communication is active. 


    Configuration

    This feature is automatically enabled by default. However, this setting can be disabled, as needed from Client Setup. Users with full access to the Client Setup permission can complete the below steps.

    1. Select All Apps in the top left corner of the header and then Settings from its drop-down menu.
    2. When the Settings page opens, select Client Setup under the Paradox Tools section.
    3. From the left panel, scroll down to and open the Employee Communications section.
    4. Toggle off Generated Summaries.
    5. Save the page.

    Important Notes:

    • Turned Off Logic: All previously generated summaries will be stored on the backend but hidden from the UI.
    • Turned Off > Turned On Again Logic: The summaries that were last generated will display, and the system will generate new summaries for Active communications that have at least 50 responses.
     

    Less than 50 responses experience

    If the communication has less than 50 responses or if the Generated Summaries setting is disabled, then the responses will display without a generated summary in a vertical scroll. When scrolling down the responses, users can select View More to load in more responses or the up arrow to return to the top of the drawer.

    • Note: You have the option of viewing an employee’s profile by clicking an employee’s name.

    50+ responses experience

    When Generated Summaries is toggled on and a communication is Active with at least 50 responses, a summary will be generated.

    The first time a user opens the Response Summary, a pop-up alongside the summary will display, notifying the user that an AI-generated summary was provided for their convenience. This is a requirement from our legal team, but once the user exits the response drawer or clicks the x icon within the popup, it won’t be shown again. This text will also display when users hover over the AI-generated summary shown below this alert.

    In the Responses section, the All tag represents all responses to the communication. When selected, the summary and responses all correspond to every response sent by an employee for that particular communication. Other tags listed after the All tag are keywords, which are described below.

    Existing Draft, Archive, Complete, and Paused communications will not have a summary generated. However, when a Paused communication is resumed, it will be treated as an Active summary and will generate a summary once it obtains 50 responses.

     

    Summary refresh

    A new summary will be generated in the following increments:

    • >100 responses = new summary every 50 responses
    • 100–300 responses = new summary every 20 responses
    • 300–1,000 responses = new summary every 50 responses
    • 1,000+ responses = new summary every 100 responses

    Keywords

    Keywords are generated based on the overall themes that the Generated Summaries feature is able to identify based on all responses to the communication. These themes are then clustered together into keyword tags. When selecting a keyword, a summary and its associated responses will display.

    A communication will never have more than 4 keywords at a time. Keywords will update over time as more responses are received.

     

    Keyword generation logic

    Each time an employee responds to a communication item, keywords are extracted from the response using an LLM. It's likely that similar keywords from different responses aren't identical, such as "management", "manager", and "manager attitude". This is when the system takes all of the unique keywords that are related to a communication item and runs them through an algorithm to sort them into separate categories. At the end of this process, similar keywords will be joined together to form one overarching keyword.

    • For example, the keywords pulled from different responses include "management", "manager", and "manager attitude". Once the related keywords are grouped together, the LLM will give them a combined label of “Management” to group those responses.

    Important Notes:

    • Keywords are extracted after every response is submitted. However, combining similar keywords into one overarching keyword takes place after at least 50 responses have been added.
    • Keywords will not be generated when communication items contain a selection type. Instead, the system will re-summarize once the distribution changes enough from the last time a summary was generated.
     

    Exporting responses

    The generated summary data will be included in the export of the communication. 


    Special case: Single-select questions

    If the communication contains a single-select communication item, a summary will display as if the sum of the messages move between percentage groups (0-10%, 11-20%, 21-30%, 31-50%, 51-70%, 71-100%). Single-select question types, including those from surveys, do not require 50+ responses.


    Employee surveys

    Employee surveys attached to an Active or Paused communication cannot be edited. This helps to reduce confusion in the summary generation of the content sent and responses received for a particular communication.

    When attempting to edit a survey tied to an Active or Paused communication, a lock icon will display next to each question, preventing the user from making updates.


    User feedback

    Users can give direct feedback on the quality of generated summaries. When a user is viewing a summary for an active communication, they will see a thumbs-up and thumbs-down icon in the bottom-right corner of the summary.

    When a user selects the thumbs-up or thumbs-down icon, the Submit Summary Feedback modal displays. The user can fill out the optional Additional details textbox to freely write their feedback.

    Once finished, the user can either select Cancel to cancel their feedback or Submit to submit their feedback.


    Permissions

    Viewing generated summaries will be based on the user’s access to the Communications permission. If a user’s role has at least view access to Communications, then they will be able to view the summary that was generated for the communication, including the keywords and keyword summaries. However, the individual responses from employees that the logged in user has access to will still be determined by their location/role/reports to permissions.

    conversational ats integrated ats frontline employee

    Was this article helpful?

    Yes
    No
    Give feedback about this article

    Related Articles

    • Conversational Career Site
    • Conversational Interview Scheduling
    • Conversational Apply

    Copyright 2026 – Paradox.

    Knowledge Base Software powered by Helpjuice

    Expand