Synthflow AI Dashboard Overview

The Dashboard Tab in Synthflow AI provides an at-a-glance overview of the key metrics and activities related to your AI-driven call campaigns. The dashboard is split into several sections, including calls, actions, and analysis. This view helps you track performance, call statistics, and the effectiveness of your assistant’s behavior.

Calls Section

The Calls section presents data on the overall number of calls and their statuses, including:

• Total Calls: Displays the total number of calls made by the assistant.
• Avg. Call Duration: Shows the average length of each call, which helps in assessing how long interactions typically last.
• Minutes Used: Provides the total minutes used in calls, useful for tracking resource consumption.

Call Status Breakdown:

The pie chart visually represents these call statuses, offering a quick overview of how calls are distributed across different outcomes.

Actions Section

The Actions section provides a summary of actions triggered during calls, such as:

• Total Actions: The total number of actions the assistant took during calls (e.g., information extraction, SMS sending).
• Appointments Scheduled: Shows how many appointments were successfully booked via the assistant.
• Live Transfers: Displays the number of calls transferred to a live agent.
• Custom Actions: Reflects any user-defined custom actions triggered during the calls.
• SMS Sent: Indicates the number of SMS messages sent by the assistant during calls.

If no actions have been set up, this section will display a message prompting the user to configure actions in the Action Tab.

Analysis Section

The Analysis section provides insights into call performance and the assistant's effectiveness in achieving predefined goals. The key metrics here include:

• Goal Achievement: Measures the percentage of calls where the assistant met the objectives outlined in its prompt (e.g., successfully booking an appointment).
• Call Drop-off Rate: Indicates how often calls were dropped before completion.
• Script Adherence: Shows whether the assistant followed the steps outlined in its prompt.
• Positive Sentiment: Reflects how positive the interaction was, measured through sentiment analysis.
• Errors: The percentage of calls that encountered issues, such as misunderstanding or technical errors.

The pie chart below this section further categorizes call endings:

• Voicemail: Calls that ended after voicemail was detected.
• Human Goodbye: Calls where the human ended the conversation.
• Undefined: Calls that ended without a specific reason being logged.

Judge Criteria Section

This section lists the Judge Criteria, which are used to evaluate the assistant’s performance during calls. It answers critical questions such as:

• Persona: Does the assistant talk about itself as defined in its prompt?
• Style: Is the tone and style consistent with what’s defined in the prompt (e.g., professional or casual)?
• No Repetition: Does the assistant avoid unnecessary repetition unless requested by the human?
• Objections Handling: Does the assistant manage objections as defined in the prompt or handle unexpected objections effectively?
• Knowledge: Does the assistant use relevant knowledge, even beyond what is specified in its prompt?
• Goal Achievement: Did the assistant achieve the call objective specified in its prompt?
• User and Agent Sentiment: Measures the positivity of both the human and the assistant during the interaction.
• Call Completion: Ensures the call ends properly with a goodbye instead of being cut off.

These criteria help ensure your assistant meets both technical and conversational goals, providing a thorough evaluation framework.

The Dashboard Tab in Synthflow AI provides an intuitive and data-rich environment for tracking and improving your assistant’s call performance, making it easy to optimize workflows and ensure high-quality customer interactions.