Admin Panel in AI Agent
Updated
The Admin Panel in AI Agent is a centralized interface that allows administrators to configure, manage, and optimize the bot’s behavior and performance. It provides intuitive controls for setting up conversation flows, voice and timeout settings, advanced AI configurations, and integration options.
Access Voicebots in AI Agents
Follow these steps to access the Admin Panel in Sprinklr AI Agent:
Open the Sprinklr Launchpad.
Navigate to AI Agent Builder.
Open the AI Agent you want to configure.
Click the Settings option in the left pane of the AI Agent application.
Expand the Voice Settings options and enable the Enable Voice toggle.
The following configurations are available in the Admin Panel:
Timeout Settings
Timeout Settings allow you to control what happens when a user does not respond within a specified time during a conversation with your AI Agent. This feature ensures the interaction remains active and prevents the conversation from stalling.
When enabled, the AI Agent can:
Wait for a defined period for the user’s reply.
Send a follow-up message or trigger another dialogue flow if no response is received.
Configuration Fields:
Timeout Settings (Toggle): Enables or disables the timeout feature.
Action:
ON: The AI Agent will monitor user inactivity and take action after the set time.
OFF: No timeout behavior will occur; the bot will wait indefinitely for a user reply.
If Timeout Settings is enabled, configure the following fields:
How long do you want to wait for a user reply?
Enter Time: Specify the numeric value (e.g., 30).
Select Unit: Choose the time unit (e.g., Seconds, Minutes).
If you enter 30 and select Seconds, the bot will wait 30 seconds for a user reply before taking action.
What reply should be sent after the timeout?
This determines what the bot does when the timeout occurs.
Enable Conditions (Toggle): Allows you to set conditions for when the timeout reply should be sent (e.g., only if the user is in a specific dialogue state).
Write Script to Publish Dynamic Asset (Toggle): Use this option to trigger dynamic content or assets instead of a static message.
Text Reply: Enter the message the bot should send after the timeout.
Example: “Are you still there? Let me know if you need help.”
(+ Add Text Reply): Use this to configure multiple follow-up messages.
Timeout Settings help maintain engagement and prevent conversations from ending abruptly. By configuring wait time and follow-up messages, you ensure a smooth and user-friendly experience.
Voice Settings and Latency Monitoring
Voice Settings define how your AI Agent listens, processes input, and delivers spoken responses during voice interactions. These settings have a direct impact on response speed, interruption handling, and overall conversation quality. To help you evaluate performance, Sprinklr records key latency metrics across speech recognition (ASR), language processing (LLM), and text‑to‑speech (TTS). These metrics allow you to understand how Voice Settings affect real‑time behavior.
Core Voice Settings
Use the following settings to control voice behavior and interruption handling.
Global Repeat Intent: Specifies the intent triggered when a user asks the bot to repeat its last response. Select an intent from the dropdown list.
Speech Profile: Defines the voice characteristics used for speech output, such as tone, language, and style. Select a predefined speech profile from the dropdown.
Accululate User Replies: Allows the voicebot to collect and combine multiple user responses into a single input before processing. This helps improve intent understanding when users speak in fragments or pause between responses.
Barge‑In Functionality: Enables users to interrupt the bot while it is speaking.
Barge‑In configuration (available when enabled):
Sensitivity (%): Minimum portion of the bot response that must play before interruption is allowed.
Time Prompt (ms): Minimum time before user interruption can occur.
Barge‑In Response Time (ms): Time taken for the bot to stop speaking after detecting user input.
Related Latency Metrics
The following metrics help you monitor performance and fine‑tune Voice Settings:
Trailing ASR Delay: Measures the time ASR takes to finalize speech after the user stops speaking.
Streamed ASR: Indicates whether ASR runs in streaming mode.
LLM TTFT / TTFS / Total LLM Time: Measures the time taken to generate AI responses.
TTS TTFB (Time to First Byte): Measures how quickly audio playback starts after a response is generated.
Total Latency: Measures the end‑to‑end time from when the user finishes speaking to when the bot begins responding.
By aligning Voice Settings with latency insights, administrators can optimize voicebot responsiveness, improve interruption handling, and deliver smoother, more natural voice interactions.
Trigger Application without Rule Setup
Allows you to launch an application without creating complex rule-based triggers.
Supported channel: Live Chat.
Select one or more accounts from the dropdown list of preconfigured accounts in Sprinklr.
Use Case Example: If you want the bot to trigger a live chat session immediately without defining conditional rules, enable this option and select the relevant account.
Feature Suite
This section enables access to additional features tailored for specific use cases. You can configure the following fields as per your requirement:
Typing Indicator (Toggle): Displays a typing indicator to users when the bot is preparing a response.
Typing Indicator Text: Customize the text shown during typing (e.g., “Bot is typing…”).
Stream Response (Toggle):
When enabled, the bot streams responses in real-time for Dynamic Workflow and Smart FAQ nodes, creating a more natural conversational experience.
Additional Settings
This section provides advanced options to fine-tune AI behavior.
AI Version: Select the model release from the dropdown to define which version generates the AI Agent’s responses.
LLM Configuration: Define which Large Language Model (LLM) configuration the bot should use for generating responses. Select the appropriate configuration from the dropdown.
The Admin Panel serves as the control center for configuring and managing your AI Agent. By using its settings you can create a personalized, efficient, and engaging user experience.