Agent Copilot Capabilities

Updated 

Once the Agent Copilot widget is configured in your Care Console, it becomes available for agents during customer interactions. The widget offers a set of AI-driven capabilities designed to improve response accuracy, reduce handling time, and assist agents in real-time. 

Note: Contact your Success Manager or Account Executive for support with enablement and configuration of Agent Copilot in your environment. 

Agent Copilot supports the following core capabilities: 

1. Case Summarization 

You can generate a real-time summary of the current case by either clicking the Summarize the case button or by typing a prompt in the message bar. This feature helps you quickly understand the case context without reading through the entire conversation. 

  • Select Add as Note to append the summary to the case as a system note. 

  • Select Save Summary to store the summary for future reference. 

2. Answer Customer Queries 

Agent Copilot uses your configured training sources—such as Knowledge Base articles, historical cases, and documents to generate accurate responses to customer queries in real time.

You can use the Answer Customer Query feature in the following ways:

  • Click the "Answer Customer Query" button
    Use the preconfigured prompt button in the Agent Copilot widget to trigger an AI-generated response.

  • Type the question directly in the message bar
    Enter a customer question manually. Agent Copilot will process the input and return a relevant response based on the available training content.

  • Use context-aware text selection in the conversation pane
    Select any part of the customer message and choose from the following options:

  • Send to Copilot: Instantly sends the selected text to Copilot and auto-generates a response.

  • Copy to Copilot: Copies the selected text to the message bar. You can optionally edit the text before submitting it to Copilot for a response.

  • Open External Knowledge Base Articles from Citations

    When you add external Knowledge Base (KB) articles to Agent Copilot as a Knowledge Source with continuous sync enabled, Copilot can now return the external article URL as a citation in its responses.

    If Copilot generates a response using content from an external KB article, the citation links directly to the original external article. Selecting the citation opens the article in the third pane, allowing you to view the source without leaving Agent Copilot.

    This feature streamlines access to external knowledge and helps you trust Copilot responses by making source attribution clear and actionable.

All responses are generated using structured and indexed training content to ensure accuracy and relevance. 

 

3. Customer Profile Overview 

You can generate a consolidated view of the customer's historical interactions and profile data to better understand the customer’s background and ongoing issues. 

  • Use the Summarse customer profile feature to access key information like contact history, preferences, and previous case details. 

 

Additional Capabilities 

Agent Copilot also includes advanced features to enhance agent productivity and optimize customer communication. 

Multilingual Support 

Agent Copilot supports multiple languages, allowing you to work in your preferred UI language. Both input and output are handled in alignment with the selected language, wherever supported. 

  • All UI elements (buttons, placeholders, response cards) appear in the selected language. 

  • Input and response text adhere to the language's direction (e.g., left-to-right or right-to-left). 

  • The ability to generate responses in the selected language depends on available training data. 

Supported languages currently include: 

Language

Code

English

en

Arabic

ar

German

de

Italian

it

French

fr

Spanish

es

 

Note: If you need to request additional languages, you must contact the Account Team or MS Team. They will then notify the product team about the requirement and use case. The product team will assess feasibility on a case-by-case basis. We support all languages supported by GPT-4o; however, for languages other than English, Arabic, German, Italian, French, and Spanish, we require a standard lead time of 3–4 weeks to fine-tune prompts and ensure optimal performance. This can be taken up by Managed Services or Implementation Consultant team. 

Feedback Mechanism

You can provide feedback on AI-generated responses to continuously improve response quality. 

 

  • Mark responses as Helpful or Not Helpful

  • For responses marked Not Helpful, provide additional feedback to assist in model retraining. 

 

Generate Email Response 

Agent Copilot can draft email responses by utilizing knowledge retrieved from training content sources. 


 

  • Add the FAQ Skill to a task to enable this functionality. 

  • The system runs a predefined FAQ flow to fetch information from KB articles, cases, or documents. 

  • Email output is structured using a specified prompt format to promote consistency. 

  • Guided Workflows (GW) are excluded from both responses and associated reporting for accuracy. 

Update Case and Profile Level Custom Field 

Agent Copilot supports updating case-level and profile-level custom fields (CFs) through natural language prompts. This functionality is powered by the UPDATE PROPERTY skill. 

To use this feature, agents can simply prompt the Copilot (e.g., “Update case priority to high”). Copilot will identify the appropriate custom fields and populate them automatically based on the provided input. 


Integrate Agent Nudges into Agent Copilot Widget

Agent Nudges are now integrated directly into the Agent Copilot widget—eliminating the need for a separate UI element in the Care Console. With this update, the Copilot becomes more proactive, offering timely prompts and best-practice reminders throughout the case lifecycle. For more details, refer to Creating Agent Nudges.


Note: This feature is in Limited Availability. Kindly reach out to your Success Manager to enable this feature in your environment. Alternatively, you can submit a request at tickets@sprinklr.com.

Translation Support for Case Summary in the Third Pane

Agent Copilot supports on-demand translation of case summaries displayed in the third pane, as well as within saved Copilot summaries. This feature enables agents to view summaries in their preferred language, regardless of the customer’s language of interaction.

This functionality is especially useful in multilingual environments, helping agents improve comprehension, follow internal workflows consistently, and respond more effectively.

How to Translate a Case Summary

Follow these steps to generate and translate a case summary:

  • Open the Case Details panel.
    The third pane will open on the right side of the screen.

  • Go to the Overview tab.

  • Under the Generate Summary section, click the Generate button.
    The system will create a case summary in the original language of the customer interaction.

  • Click the Original language button in the summary header.
    A dropdown will appear with supported languages.

  • Select your preferred language from the dropdown.
    The translated summary will be generated instantly.

  • Click 'Edit as a Note' to edit the translated summary.

    This will opens the translated summary in the reply box (publisher) where you can edit the text.

    Upon submission, the content is logged as a note (not a message) in the case conversation panel.

  • Click 'Add as a Note' to save the translated summary.
    This will save the translation as a system note within the case without editing.

Note: Translations generated in the third pane are not saved automatically. You must click Add as a Note to retain the translation for future reference.

Agent Copilot equips agents with intelligent, real-time assistance to streamline case handling, improve response accuracy, and enhance overall customer experience. With capabilities like case summarization, query resolution, customer profile insights, and multilingual support, Agent Copilot helps reduce average handling time and boosts agent efficiency across service interactions.