Call Sentiment Analysis Workflow in Kylas CRM

The Call Sentiment Analysis Workflow in Kylas CRM leverages AI to automatically analyze the tone, emotion, and sentiment of recorded calls between your agents and customers.

This helps sales and service teams understand not just what was said, but how it was said — empowering managers to assess customer satisfaction, agent empathy, and overall communication quality in real time.

Table of Contents

1. Overview

2. Prerequisites

3. How It Works

4. Steps to Set Up a Call Sentiment Analysis Workflow

   • Step 1: Create Workflow for Call Logs

   • Step 2: Define Trigger & Conditions

   • Step 3: Add “Generate Insight” Action

   • Step 4: Configure Your AI Prompt for Sentiment Analysis

   • Step 5: Save and Activate

5. Example AI Prompts

6. Example Output

7. Use Cases

8. Benefits

9. Editable & Customizable Workflow

10. Conclusion

Overview

The Sentiment Analysis Workflow is a powerful automation within Kylas CRM’s workflow engine that uses AI to evaluate recorded conversations automatically.

It listens to call recordings, analyzes language, tone, and keywords, and determines whether the customer’s overall experience was positive, neutral, or negative — along with reasons and observations.

This helps businesses monitor customer satisfaction, identify training opportunities, and ensure consistent call quality without manually reviewing every recording.

Prerequisites

Before setting up the Sentiment Analysis workflow, ensure you have:

• ✅ A Kylas CRM account with Workflow Automation enabled

• ✅ Access to Call Logs and AI Actions (Sherpa AI Beta)

• ✅ Call recordings being captured via your telephony integration

• ✅ Workflow permission to use Generate Insight or AI-based actions

How It Works

1. Every time a call is completed and logged in Kylas CRM,

2. The workflow automatically checks conditions like call duration and recording availability,

3. The AI engine then analyzes tone, language, and emotion of both parties,

4. Generates a sentiment score (Positive / Neutral / Negative) and AI summary of emotional flow,

5. Results are added to the Call Log as Sentiment Insight and Notes.

Steps to Set Up Call Sentiment Analysis Workflow

Step 1: Create Workflow for Call Logs

1. Navigate to Settings → Automation → Workflows.

2. Click + Create New Workflow.

3. Name it — for example:
   “Call Sentiment Analysis Workflow”.

4. Under Select Entity, choose Call Log.

Step 2: Define Trigger and Conditions

• Trigger Type: When a Call Log is Created.

• Trigger Preference: Immediate Action (so it runs as soon as the call is logged).

Then, add conditions:

Condition	Operator	Value
Duration (In Seconds)	Greater Than	20
Recording File	Is Set	—

💡 This ensures that only valid calls with audio files longer than 20 seconds are analyzed.

Step 3: Add “Generate Insight” Action

1. Under Set Actions to be Performed, select:
   Action Type → Generate Insight

2. This allows you to use AI to interpret and evaluate call emotion.

Step 4: Configure Your AI Prompt for Sentiment Analysis

In the prompt text box, enter your AI instructions — explaining how the call’s emotion and tone should be analyzed.

Here’s an example you can use 👇

Example Prompt

You are an AI assistant that analyzes the tone and emotion of recorded sales or service calls. 
Listen to the entire conversation between the agent and the customer and evaluate the sentiment.

Steps to follow:
1. Identify the emotional tone of the customer (positive, neutral, negative).
2. Identify the tone of the agent (calm, persuasive, rude, helpful, empathetic, etc.).
3. Detect changes in tone or mood during the call (for example, frustrated at the beginning, satisfied at the end).
4. Provide a short overall sentiment summary and sentiment rating:
   - Positive 😊
   - Neutral 😐
   - Negative 😠
5. Mention 2-3 sentences explaining why this rating was given.
6. End with recommendations for the agent (if applicable).

Step 5: Save and Activate the Workflow

• Click Save once the workflow setup is complete.

• Then click Activate.

From now on, every time a call meeting your criteria is logged, the workflow will automatically analyze sentiment and log AI-generated feedback inside the Call Record.

Example Output

Here’s an example of what the AI Sentiment Insight might look like inside your CRM call record 👇

Call Sentiment Summary:

Overall Sentiment: 😊 Positive

Customer Tone: Initially skeptical but became more engaged and satisfied after the agent explained the loan details clearly.

Agent Tone: Calm, professional, and empathetic throughout the conversation.

Observation:

• Customer was frustrated about the high existing interest rate (00:15–01:12).

• Agent handled objections patiently and offered clear solutions (01:30–03:20).

• Customer agreed to proceed by end of the call (04:10–04:45).

Recommendation:
Maintain the same calm tone and continue focusing on active listening and reassurance.

Use Cases

Use Case	Description
Customer Experience Monitoring	Evaluate customer satisfaction across multiple calls automatically.
Sales Performance Tracking	Measure how effectively agents handle objections and build trust.
Quality Audits	Automate QA reviews for tone and empathy compliance.
Escalation Alerts	Identify calls with negative sentiment for quick review.
Team Coaching	Use sentiment summaries for performance feedback and training.

Benefits

Benefit	Description
💬 Emotional Insight	Understand customer and agent emotions, not just words.
⚙️ Automated Evaluation	AI handles tone analysis for every call — no manual listening.
🔁 Consistent QA	Standardized tone scoring across all teams.
📈 Improved Training	Helps managers coach agents using real examples.
⏱️ Time-Saving	Cuts down hours spent manually reviewing calls.
💡 Better CX Tracking	Identify patterns of dissatisfaction early.

Editable & Customizable Workflow

The best part — this workflow is completely editable.

You can:

• Modify the AI prompt (sales, support, healthcare, insurance, etc.)

• Add or remove questions such as:

  • “Was the customer frustrated?”

  • “Did the agent sound empathetic?”

  • “How did the customer tone change through the call?”

• Clone this workflow for multiple teams (Sales, Support, Collections, etc.)

• Set different duration thresholds for call length filtering

This flexibility ensures sentiment analysis adapts perfectly to your business process.

Conclusion

The Call Sentiment Analysis Workflow in Kylas CRM helps businesses go beyond traditional call summaries — enabling emotional intelligence at scale.

By automating tone evaluation, managers can instantly know how well agents are communicating, how customers feel, and where improvement is needed.

Together with Call Insight Workflows, this forms a powerful combination — one that tells you both:

• 🗒️ What happened in the call

• 💬 How it went emotionally

Use this workflow to enhance call quality, elevate customer satisfaction, and build smarter, empathetic sales and service teams inside Kylas CRM.