🎧 Generate Call Insights by Adding Custom Questions for Call Transcripts via Workflow in Kylas CRM

Kylas CRM’s AI-powered call summary and insight generation workflow allows you to automatically analyze and summarize recorded calls — helping teams extract meaningful insights such as customer intent, agent performance, product discussions, and follow-up actions.

This feature leverages AI-generated call transcripts and custom business questions that you can configure in a workflow — so every call recording is analyzed as per your business context (e.g., sales, support, banking, healthcare, etc.).

Table of Contents

1. Overview

2. Prerequisites

3. How It Works

4. Steps to Set Up Call Transcript and Insight Workflow

   • Step 1: Access Workflows

   • Step 2: Create a Workflow for Call Logs

   • Step 3: Define Trigger and Conditions

   • Step 4: Add the “Generate Call Summary” Action

   • Step 5: Add Custom Questions to Generate Insights

   • Step 6: Save and Activate

5. Example Prompt and Questions

6. Understanding Generated Insights

7. Use Cases

8. Benefits

9. Editable & Customizable Setup

10. Conclusion

 Overview

The Call Sentiment Analysis & Insight Workflow in Kylas CRM allows businesses to:

• Automatically transcribe recorded calls using AI.

• Generate a structured call summary in a timestamped format.

• Add custom questions to extract insights specific to your business operations.

• Automatically populate these summaries and insights into the Call Log Notes and Insights Section of the CRM.

This ensures faster, more accurate post-call reviews — without the need for manual listening or note-taking.

Prerequisites

Before setting up the workflow, ensure you have:

• A Kylas CRM account with Workflow Automation enabled.

• Access to the Call Logs entity.

• Active call recordings being logged in the system.

• AI-powered workflow permissions (Sherpa AI Beta / AI Action enabled).

How It Works

1. When a call is logged in CRM (e.g., from integrated telephony),

2. The workflow automatically checks if the call meets the defined criteria (like duration, recording availability).

3. The system uses AI to transcribe the recording based on a custom prompt.

4. The transcription and answers to predefined questions are generated and added to the call record.

Example Output:

• Timestamped transcript

• Summary

• Insights answering key questions (like “Did the agent explain the product properly?”)

Steps to Set Up Call Transcript & Insight Workflow

Step 1: Access Workflows

1. Go to Settings → Automation → Workflows.

2. Click + Create New Workflow.

3. Give your workflow a name — e.g., Call Sentiment Analysis Workflow.

Step 2: Create a Workflow for Call Logs

• Select Entity: Choose Call Log.

• This ensures the workflow runs automatically when a call log is created.

Step 3: Define Trigger and Conditions

1. Trigger Event:

   • Choose When a call log is created.

   • Select Immediate Action to process as soon as the call is logged.

2. Set Conditions:

   • Choose Based on conditions.

   • Example setup:

     • Duration (In Seconds) → Greater Than → 20

     • Recording File → Is Set

📌 This ensures that only valid, recorded calls longer than 20 seconds are analyzed.

Step 4: Add the “Generate Call Summary” Action

1. Under Set actions to be performed, select Generate Call Summary.

2. Enter your AI Prompt that defines how the transcription should be structured.

3. Example Prompt:

   You are a helpful assistant that summarizes doctor-patient consultations. 
   Given an audio clip of a medical consultation between a healthcare provider and a patient, 
   transcribe the conversation in the same language as the audio. 
   Format the transcription in a timestamped dialogue style like this:
   
   Example:
   00:00 Speaker A: Good morning. What brings you in today?
   00:04 Speaker B: I've been having chest pains for the past week.
   
   If actual names are not available, use placeholders like Speaker A, Speaker B, etc.

💡 This prompt can be edited as per your business needs — for sales, insurance, or support conversations.

Step 5: Add Custom Questions to Generate Insights

Below the prompt section, click on Generate Insight → Add Questions.
Add the questions you want AI to answer based on the transcript.

Example Questions for a Banking Sales Call:

1. Was proper product knowledge shared by the agent?

2. Did the agent ask for the customer’s phone number?

3. What is the customer’s loan amount, CIBIL score, and current bank?

4. What was the customer’s exact requirement?

5. What action did the agent take after the discussion?

AI will automatically extract and summarize answers from the call transcript.

Step 6: Save and Activate

• Click Save to finalize your workflow setup.

• Click Activate to start automating call transcription and insight generation.

Once active, every call that meets the set conditions will be automatically transcribed and analyzed.

Example Output: Generated Insights

Below is an example of how the insights appear in the CRM after the workflow runs:

Call Insights (Example: Home Loan Call)

1️⃣ Given proper product knowledge?
Yes, the agent demonstrated product knowledge by explaining the Central Bank of India’s home loan balance transfer offer at a 7.35% rate (00:00-00:01).

2️⃣ Agent asked for customer number?
Information not available in transcript.

3️⃣ Location, volume, CIBIL, monthly income, product, current bank, existing interest rate?

• Location: Kudla West (02:33)

• Volume: ₹73 lakh outstanding (01:27)

• CIBIL: Not confirmed by customer (02:26-02:28)

• Monthly Income: ₹35,000 (03:16)

• Product: Home Loan Balance Transfer

• Current Bank: Tata Capital

• Existing Rate: 8.25% (02:30)

4️⃣ What is the exact requirement?
Customer wanted to lower the home loan interest rate via balance transfer.

5️⃣ Agent Action Taken?
Agent offered a 7.35% Central Bank balance transfer, collected loan details, and scheduled a callback (02:37–04:54).

Use Cases

Use Case	Description
Sales Quality Analysis	Review whether agents followed the pitch correctly.
Customer Service Monitoring	Identify tone and sentiment in customer support calls.
Healthcare Calls	Auto-generate medical summaries for doctor-patient interactions.
Compliance Audits	Ensure agents meet communication and disclosure standards.
Training Insights	Use AI-generated reports to coach teams effectively.

Benefits

Benefit	Description
🔍 Automated Transcription	No manual effort — get timestamped, speaker-tagged transcriptions.
🤖 Customizable Prompts	Modify workflow prompts as per business type — healthcare, BFSI, sales, etc.
📊 Actionable Insights	Automatically answer domain-specific questions for deeper analysis.
⚡ Improved Productivity	Save hours spent listening to calls manually.
📁 Centralized Data	Store call summaries and insights directly inside each Call Log.
🧠 Continuous Learning	Review agent patterns, improve training, and boost call quality.

Editable & Customizable Setup

The Generate Call Summary workflow is fully editable — you can:

• Modify or replace the AI prompt anytime.

• Add or remove custom questions as per evolving business needs.

• Clone existing workflows for different departments (Sales, Support, Healthcare, etc.).

• Adjust conditions (e.g., duration > 10s or include missed calls).

This flexibility allows every business to create AI workflows tailored to its call-handling process.

Conclusion

The Generate Call Insights Workflow in Kylas CRM combines the power of AI-driven transcription with custom business intelligence.

By setting up the right prompts and questions, you can transform ordinary call logs into structured reports — capturing product knowledge, customer needs, agent performance, and next-step actions effortlessly.

This feature is perfect for businesses that rely on phone-based sales, service, or consultations — empowering them to analyze conversations, improve outcomes, and enhance customer experience at scale.