Help Center

AI Phone Call Sentiment Analysis

Call Sentiment Analysis uses AI to review your recorded phone calls and evaluate the emotional tone of the conversation. It identifies whether the call felt positive, neutral, or showed concern, helping you quickly understand how the interaction went without re-listening to the call.

This feature is useful for:

• Quickly identifying calls where a client may be unhappy or concerned
• Reviewing successful conversations to understand what worked well
• Getting a high-level emotional read without replaying the recording
• Tracking client sentiment trends over time

Requirements

Before sentiment analysis can be generated:

• The call must be recorded
• The call must be transcribed

If a call is not transcribed, the AI will not be able to analyze sentiment.

How to Generate Call Sentiment Analysis

Sentiment analysis can be generated automatically or manually.

Automatic Sentiment Analysis (Recommended)

When enabled, sentiment analysis is generated automatically for every recorded and transcribed call.

1. Click your profile menu in the top navigation bar.
2. Select Settings.
3. Go to Phones and texts → Phone numbers.
4. Click the Settings (gear icon).
5. Select Claudia AI settings.
6. Enable Call sentiment analysis.
7. Click Save.

Once enabled, sentiment analysis will be generated automatically for all future calls on that number.

Manual Sentiment Analysis (On-Demand)

You can also generate sentiment analysis for an individual call.

1. Go to Leads → Calls.
2. Find the recorded call you want to analyze.
3. Click the three-dot menu (⋮) next to the call.
4. Select Generate sentiment analysis.

The AI will begin analyzing the call immediately.

Understanding Sentiment Results

The sentiment analysis provides several types of insights.

Overall Sentiment

A percentage-based breakdown showing how the call was perceived overall:

• Positive — Parts of the call that reflected satisfaction, confidence, or enthusiasm
• Neutral — Informational or matter-of-fact sections of the conversation
• Concern — Areas where frustration, worry, or dissatisfaction was detected

Sentiment by Topic

Sentiment is analyzed across different topics discussed during the call. For example, a call may show positive sentiment overall, but higher concern when discussing fees or timelines.

Key Moments

The AI highlights specific moments that influenced sentiment, including:

• Positive moments — When the client expressed approval or satisfaction
• Concern moments — When the client raised objections, concerns, or frustrations

Where to Find Sentiment Analysis Results

After analysis is complete (usually within 1–2 minutes):

• You will receive an email notification
• A sentiment analysis note is added to the related contact record

The note is linked to the specific call and saved alongside other call-related notes.

The sentiment analysis note includes:

• An overall sentiment breakdown
• Context explaining what influenced the sentiment
• Highlighted moments that affected the tone of the call

Using Sentiment Analysis Effectively

Identify At-Risk Clients Early

Calls with a high level of concern may indicate issues that need immediate follow-up. Addressing them quickly can help prevent escalation.

Learn From High-Performing Calls

Calls with strong positive sentiment can help identify effective communication patterns and techniques worth repeating.

Track Long-Term Trends

Over time, sentiment analysis can reveal patterns related to specific topics, advisors, or call types.

Troubleshooting

• “Call transcription not found”
  The call must be transcribed before sentiment analysis can be generated.
• Sentiment option not available
  Make sure the call has an audio recording. Only recorded calls can be analyzed.
• Results seem inaccurate
  Sentiment is inferred from text. Sarcasm, humor, or cultural nuances may affect accuracy. Use sentiment analysis as one data point, not a final judgment.