AI Receptionist (Inbound Calls via Voice AI)

General Overview

AI Receptionist is a type of AI agent in Kayse designed to handle inbound calls. The system automatically answers incoming calls, analyzes the caller's speech, performs required actions (e.g., transferring to a live agent, creating a task, sending document upload links), and logs the entire interaction for analytics.

Main Interface Blocks & Functionality

1. Metrics Block Components:

• Total Calls — total number of inbound calls within the selected date range.
• Time on Calls — cumulative duration of all calls.
• Avg. Call Time — average duration per call.
• New Leads — number of new clients acquired through calls.
• Avg. Sentiment — average emotional sentiment (Happy / Neutral / Unhappy).

Metric Comparison with Previous Period

• All top-level metrics (Total Calls, Avg. Call Time, etc.) are compared with the previous period of the same length.
• For example:
  • Past 7 days (June 9–15) → compared with June 2–8
  • Past 30 days → compared with the 30 days before that
  • This month → compared with previous month (if "previous period" logic is applied)
• If the previous period had 0 calls, the increase is shown as 100%, but should be interpreted carefully.

2. Calls Past Block Functionality:

• Displays call counts over time.
• Categories: Missed, Transferred, Voicemail, Answered.
• Visualizes call activity trends.

Interpretation of Call Statuses

3. Actions Taken Functionality:

• Bar chart showing the actions performed after calls:
  • 1. Qualified – Intake Completed

✅ The client successfully completed the intake process either through the AI Receptionist.

This status indicates a positive outcome where all required information was collected, and the case is ready for legal review or assignment.

• 2. Transferred to Agent

🔄 The AI Receptionist determined that the call required human assistance and successfully transferred the caller to a live agent or a different department.

• 3. Voicemail Taken

📩 A voicemail message was recorded from the caller—either automatically after timeout or if the user chose not to speak directly.

• 4. Missed Calls

❌ The call was not handled by the AI Receptionist or a live agent. This typically happens due to reaching the concurrency limit—when all available AI agents are already handling other calls.

4. Call Success Rate by Day/Time Functionality:

This section displays a heatmap that visualizes the success rate of calls based on the day of the week and time of day.

Functionality:

• Each cell represents a time window (2-hour block) for a specific day.
• The percentage value inside each cell shows the call success rate—i.e., how many calls during that time slot resulted in a successful conversation with a client (either via AI agent or live agent).

Color intensity indicates performance:

• Dark Red = low success rate
• Bright Blue = high success rate
• Gradient shades help quickly identify strong and weak time slots.

5. Recent Calls Functionality:

This section presents a real-time chronological list of the most recent inbound or outbound calls handled by the AI Receptionist system.

Functionality:

• Displays key call details at a glance:

  • Caller Name (if available) or New Caller + phone number
  • Call duration
  • Timestamp (date & time)

• Supports quick visibility into daily call traffic and immediate quality monitoring.

Color-Coded Status Indicators:

Each call entry is accompanied by a colored icon representing its outcome:

• 🟢 Green — Successful Call: The call was answered and a conversation took place.
• 🟡 Yellow — Missed Call: The system reached its concurrency limit and couldn’t handle the call.
• 🔵 Blue — Transferred Call: The AI transferred the call to a live agent.
• ⚪ Gray — Voicemail: (Currently unused in AI Receptionist flows, but reserved for future logic if implemented).

6. Calls Tab (Detailed Call View) Functionality:

This tab provides an in-depth overview of each individual call handled by the AI Receptionist, allowing operators, QA teams, or managers to review call performance, verify outcomes, and track agent-client interactions in a structured and actionable format.

Functionality:

The call view is divided into three clear panels:

A. Call List Panel (Left Sidebar)

• Displays a scrollable list of recent calls with:

  • Caller Name
  • Call Duration
  • Timestamp (date & time)
  • Status indicator (color-coded)

• Supports searching by name or number for easy access.

🔍 If the caller's name is not available, the system displays fallback formatting like “New Caller + Phone Number”.

B. Call Summary Panel (Middle)

Shows key metadata for the selected call, including:

• Sentiment: AI-detected mood (e.g., Happy, Neutral, Unhappy)

• Call Result: e.g., Successful, Missed, Transferred

• Disconnection Reason: e.g., User Hangup, Agent Ended, Busy

• Case Information:

  • Case Name
  • Case Type
  • Case Number

Provides a quick glance into the context and classification of the interaction.

C. Transcript & Audio Panel (Right)

• Full text transcript of the call conversation, broken down by speaker (Agent vs. Client).

• Ability to search within the transcript to locate key phrases.

• Audio recording player with controls:

  • Play/pause
  • Adjust playback speed (1x, 1.5x, 2x)
  • Download option for offline review

• At the top, a summary box highlights the purpose and outcome of the call (auto-generated by AI).

Date Range Logic