Agent Copilot Additional Capabilities
Updated
Agent Copilot offers enhanced features to improve accuracy and efficiency when assisting agents during customer interactions.
Select Specific Conversation Context for Copilot Actions
Agents can now choose specific parts of a conversation transcript or speech-to-text (STT) input within the Care Console and use them as context for Copilot actions. This ensures that summaries, answers, and recommendations are based only on relevant conversation segments.
Steps to Send Selected Messages to Agent Copilot:
In the Care Console conversation view, select the first message you want to include.
Press and hold Shift, then select the last message in the range. All messages between the first and last will be selected automatically.
A bar appears at the bottom of the window. Click the ellipsis (⋯) icon and select Add to Copilot.
The selected conversation segment is added to Copilot.
Perform the desired action in Copilot. Only the selected messages will be considered.
Tip: You can create multiple groupings from the same conversation within a case.
Benefits:
Improved Accuracy: Agents can summarize or process conversations more precisely when multiple topics are discussed in a single case.
Reduced Noise: Focuses on relevant context, eliminating confusion caused by unrelated information.
Increased Efficiency: Prevents errors caused by applying workflows to the wrong context, saving time and effort.
Streaming Support in Agent Copilot
Agent Copilot now supports streaming responses, delivering content progressively in the UI as it is generated. This enhancement improves reliability and provides a smoother, more responsive experience for agents.
Note: Access to this feature is controlled by a dynamic property (DP). To enable this feature in your environment, reach out to your Success Manager. Alternatively, you can submit a request at tickets@sprinklr.com.
What Is Streaming?
Streaming allows agents to start reading and using the response without waiting for the entire output to load. This is particularly useful for longer outputs such as:
Case summaries
Detailed answers
Complex recommendations
By reducing perceived response time, streaming improves overall efficiency and responsiveness during customer interactions.
Key Benefits
Improved Reliability: Ensures consistent response delivery even for large outputs.
Faster Access to Information: Agents can act on partial responses immediately.
Enhanced User Experience: Content appears progressively, reducing wait frustration.
Note:
Starting with release 26.4.1, streaming is supported directly for the FAQ function. This update does not affect proactive or scheduled FAQs.
Select and Copy Copilot’s Response
Agent Copilot now allows agents to select specific sections of its response and take targeted actions, improving flexibility and reducing manual effort.
Follow the steps below to use this capability:
Highlight the desired portion of Copilot’s response.
Right-click on the selected text.
Choose one of the following options:
Copy to Copilot: Use the selected text to refine a follow-up query.
Add to Response: Insert the selected text directly into a customer reply.
Only the highlighted text is carried forward, ensuring that agents work with the most relevant content.
.png)
Benefits
Time Savings: Reduces extra clicks and manual copy-paste actions.
Improved Accuracy: Ensures only relevant content is included in follow-up queries or customer responses.
Enhanced Efficiency: Streamlines workflows during customer interactions.
Proactive Upsell Recommendations
Agent Copilot can proactively identify upsell opportunities during live customer conversations and surface them as an Upsell card within the Copilot widget. When a configured upsell scenario is detected, Copilot presents a contextual recommendation along with a suggested response that agents can review before engaging with the customer.
How It Works
When the conditions defined for an upsell recommendation are met, Agent Copilot automatically displays an Upsell card during the conversation. The card provides:
- A recommended upsell opportunity based on the conversation context.
- A pre-generated message draft that agents can review and refine before sending.
- Action options that allow agents to proceed with the recommendation or dismiss it.

Benefits
This capability helps agents:
- Identify revenue opportunities without manually analyzing the conversation.
- Deliver relevant offers at the right moment in the customer journey.
- Maintain a natural conversation flow while promoting additional products or services.
- Improve consistency in upsell messaging through AI-generated recommendations.
Administrators can configure proactive tasks to generate Upsell cards and define the prompts used to create upsell recommendations. When a task is configured with the Upsell card type, Agent Copilot renders the recommendation using the Upsell card experience in the Copilot widget.
Agent Experience
When an Upsell card appears:
- Review the recommended offer.
- Review and edit the suggested message, if needed.
- Share the recommendation with the customer or dismiss the suggestion.
- Continue the conversation without switching between tools or workflows.
Note: Upsell recommendations are surfaced only when the configured proactive task conditions are met during a customer interaction.
This feature enables organizations to embed intelligent sales guidance directly into customer conversations, helping agents identify and act on upsell opportunities more effectively.
System Notifications for Proactive Copilot Suggestions
System Notifications help ensure that agents never miss important Agent Copilot recommendations during active customer interactions. When Agent Copilot generates a proactive suggestion, Sprinklr can display a system-level notification that alerts agents immediately, even when they are working in another browser tab or application.
How It Works
Agent Copilot continuously analyzes customer conversations and surfaces proactive recommendations, such as:
- Upsell opportunities
- Sales recommendations
- Promise Management suggestions
- Escalation recommendations
- Other configured proactive Copilot tasks
When a proactive recommendation is generated, Sprinklr automatically sends a system notification to the agent.

Receive Notifications Outside Sprinklr
System notifications remain visible even when:
- Sprinklr is running in a background browser tab.
- Another application is currently active.
- The agent is not actively viewing the Care Console.
Notifications use supported browser and operating system notification capabilities to ensure proactive recommendations remain visible throughout the interaction.
Open Recommendations Directly from a Notification
When an agent selects a system notification, Sprinklr automatically opens the associated interaction and navigates the agent directly to the relevant Copilot recommendation. This helps agents review and act on recommendations without manually locating the conversation or suggestion.
Administrators can control the availability of proactive Copilot notifications and determine whether agents receive system notifications for proactive recommendations.
Notification delivery also respects:
- Browser notification permissions
- Operating system notification settings
- User notification preferences
Agents must grant the necessary notification permissions to receive system notifications.
Benefits
System Notifications help organizations:
- Improve visibility of real-time Copilot recommendations.
- Reduce the likelihood of missed proactive suggestions.
- Enable faster responses to time-sensitive guidance.
- Increase adoption of Agent Copilot recommendations.
- Help agents remain productive while working across multiple applications.
Note: System notifications are generated only when Agent Copilot surfaces a proactive recommendation during an active customer interaction. Notifications support all proactive Copilot card types and are subject to browser, operating system, and platform notification permissions.
These capabilities in Agent Copilot such as selective copying, dynamic query refinement, and streamlined response insertion enable agents to work more efficiently and accurately. By using these capabilities, agents can reduce manual effort, improve response quality, and deliver faster, more personalized customer interactions.