Sales Call Agent

The Sales Call Agent provides automated analysis of sales call recordings. It transcribes audio files and generates comprehensive insights including client sentiment, products mentioned, pricing discussions, action items, and conversation summaries. This agent helps sales teams quickly review and analyze their calls for better follow-up and performance tracking.

Base URL

/api/agents/sales_agent

Authentication

All endpoints require authentication. Sign up to the https://nextneural.superteams.ai to get your API key.

How It Works

The Sales Call Agent performs comprehensive sales call analysis:

Audio Transcription: Converts sales call audio to text using AI-powered speech recognition
Intelligent Analysis: Automatically extracts:
- Speaker identification (Sales Rep vs Client)
- Client sentiment analysis (Positive, Negative, Neutral)
- Products and services mentioned
- Pricing discussions and mentions
- Action items and commitments from both parties
- Conversation summary
Caching: Automatically caches results to avoid re-processing the same audio
Re-evaluation: Supports force re-processing for updated analysis

Endpoints

1. Health Check

Check if the Sales Call Agent service is running.

Endpoint: GET /health

Authentication: None required

Response:

{
  "status": "healthy",
  "service": "sales_call_agent"
}

2. Analyze Sales Call

Process a sales call audio file and generate comprehensive analysis.

Endpoint: POST /analyze_call

Authentication: Required

Request Body:

{
  "file_name": "sales_call_2025_01_14.mp3",
  "document_id": 123,
  "force_reprocess": false
}

Parameters:

file_name (required): Filename of the audio in the media directory
document_id (required): Reference to the knowledge base document ID for ownership verification
force_reprocess (optional, default: false): Re-analyze call even if already processed

Request Example:

curl -X POST "https://nextneural-api.superteams.ai/api/agents/sales_agent/analyze_call" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "file_name": "sales_call_2025_01_14.mp3",
    "document_id": 123,
    "force_reprocess": false
  }'

Response (New Analysis):

{
  "analysis": {
    "id": 456,
    "userId": 1,
    "kbDocumentId": 123,
    "audioFilename": "sales_call_2025_01_14.mp3",
    "callDuration": "00:15:30",
    "clientSentiment": "Positive",
    "productsMentioned": [
      "Enterprise Plan",
      "Premium Support Package",
      "API Integration Module"
    ],
    "pricingMentions": [
      "$5,000 per month for Enterprise Plan",
      "20% off for annual commitment",
      "$500 setup fee waived for early adopters"
    ],
    "actionItems": [
      "Sales Rep will send a detailed pricing proposal by end of day Friday",
      "Client will review the product demo video and share feedback by next Tuesday",
      "Sales Rep will schedule a follow-up call with the technical team next week",
      "Client will connect Sales Rep with the procurement department for contract discussion"
    ],
    "conversationSummary": "Client expressed strong interest in the Enterprise Plan with API integration. Discussed pricing options and agreed on next steps for technical evaluation and procurement process.",
    "transcript": [
      {
        "speaker": "Sales_Rep",
        "text": "Good morning! Thank you for taking the time to speak with me today. I understand you're interested in our Enterprise Plan."
      },
      {
        "speaker": "Client",
        "text": "Yes, we've been evaluating several solutions and your platform looks promising for our needs."
      }
    ],
    "alreadyEvaluated": false,
    "evaluatedAt": "2025-01-14T10:30:00",
    "isReevaluation": false,
    "createdAt": "2025-01-14T10:30:00",
    "updatedAt": "2025-01-14T10:30:00"
  }
}

Response (Cached Result):

{
  "analysis": {
    "id": 456,
    "userId": 1,
    "kbDocumentId": 123,
    "audioFilename": "sales_call_2025_01_14.mp3",
    "callDuration": "00:15:30",
    "clientSentiment": "Positive",
    "productsMentioned": [...],
    "pricingMentions": [...],
    "actionItems": [...],
    "conversationSummary": "...",
    "transcript": [...],
    "alreadyEvaluated": true,
    "evaluatedAt": "2025-01-14T10:30:00",
    "isReevaluation": false,
    "createdAt": "2025-01-14T10:30:00",
    "updatedAt": "2025-01-14T10:30:00"
  }
}

Response (Re-evaluation):

{
  "analysis": {
    "id": 456,
    "userId": 1,
    "kbDocumentId": 123,
    "audioFilename": "sales_call_2025_01_14.mp3",
    "callDuration": "00:15:30",
    "clientSentiment": "Positive",
    "productsMentioned": [...],
    "pricingMentions": [...],
    "actionItems": [...],
    "conversationSummary": "...",
    "transcript": [...],
    "alreadyEvaluated": false,
    "evaluatedAt": "2025-01-14T10:30:00",
    "isReevaluation": true,
    "createdAt": "2025-01-14T10:30:00",
    "updatedAt": "2025-01-14T11:45:00"
  }
}

Notes:

The audio file must exist in the configured media directory
Document ownership is verified before processing
If force_reprocess=false and analysis exists, returns cached result
If force_reprocess=true, re-analyzes and updates existing record
High-accuracy AI transcription and analysis ensures quality results
Each document can only have one analysis record

3. Get My Analyses

Retrieve all sales call analyses for the authenticated user.

Endpoint: GET /my_analyses

Authentication: Required

Query Parameters:

limit (optional, default: 100): Maximum number of records to return

Request Example:

curl -X GET "https://nextneural-api.superteams.ai/api/agents/sales_agent/my_analyses?limit=50" \
  -H "Authorization: Bearer YOUR_TOKEN"

Response:

{
  "success": true,
  "userId": 1,
  "count": 3,
  "analyses": [
    {
      "id": 456,
      "userId": 1,
      "kbDocumentId": 123,
      "audioFilename": "sales_call_2025_01_14.mp3",
      "callDuration": "00:15:30",
      "clientSentiment": "Positive",
      "productsMentioned": ["Enterprise Plan", "API Integration"],
      "pricingMentions": ["$5,000 per month"],
      "actionItems": [
        "Sales Rep will send pricing proposal by Friday",
        "Client will review demo video by Tuesday"
      ],
      "conversationSummary": "Client interested in Enterprise Plan...",
      "transcript": [...],
      "alreadyEvaluated": false,
      "evaluatedAt": "2025-01-14T10:30:00",
      "isReevaluation": false,
      "createdAt": "2025-01-14T10:30:00",
      "updatedAt": "2025-01-14T10:30:00"
    },
    {
      "id": 455,
      "userId": 1,
      "kbDocumentId": 122,
      "audioFilename": "sales_call_2025_01_13.mp3",
      "callDuration": "00:22:15",
      "clientSentiment": "Neutral",
      "productsMentioned": ["Standard Plan"],
      "pricingMentions": ["$2,000 per month"],
      "actionItems": [
        "Sales Rep will provide case studies by next week"
      ],
      "conversationSummary": "Initial discovery call...",
      "transcript": [...],
      "alreadyEvaluated": false,
      "evaluatedAt": "2025-01-13T15:20:00",
      "isReevaluation": false,
      "createdAt": "2025-01-13T15:20:00",
      "updatedAt": "2025-01-13T15:20:00"
    }
  ]
}

Notes:

Returns analyses in reverse chronological order (newest first)
Only returns analyses belonging to the authenticated user
Includes full analysis details for each record

4. Get Specific Analysis

Retrieve detailed information for a specific sales call analysis.

Endpoint: GET /analysis/{record_id}

Authentication: Required

Path Parameters:

record_id (required): The ID of the analysis record to retrieve

Request Example:

curl -X GET "https://nextneural-api.superteams.ai/api/agents/sales_agent/analysis/456" \
  -H "Authorization: Bearer YOUR_TOKEN"

Response:

{
  "success": true,
  "analysis": {
    "id": 456,
    "userId": 1,
    "kbDocumentId": 123,
    "audioFilename": "sales_call_2025_01_14.mp3",
    "callDuration": "00:15:30",
    "clientSentiment": "Positive",
    "productsMentioned": [
      "Enterprise Plan",
      "Premium Support Package",
      "API Integration Module"
    ],
    "pricingMentions": [
      "$5,000 per month for Enterprise Plan",
      "20% off for annual commitment",
      "$500 setup fee waived for early adopters"
    ],
    "actionItems": [
      "Sales Rep will send a detailed pricing proposal by end of day Friday",
      "Client will review the product demo video and share feedback by next Tuesday",
      "Sales Rep will schedule a follow-up call with the technical team next week",
      "Client will connect Sales Rep with the procurement department for contract discussion"
    ],
    "conversationSummary": "Client expressed strong interest in the Enterprise Plan with API integration. Discussed pricing options and agreed on next steps for technical evaluation and procurement process.",
    "transcript": [
      {
        "speaker": "Sales_Rep",
        "text": "Good morning! Thank you for taking the time to speak with me today."
      },
      {
        "speaker": "Client",
        "text": "Yes, we've been evaluating several solutions."
      }
    ],
    "alreadyEvaluated": false,
    "evaluatedAt": "2025-01-14T10:30:00",
    "isReevaluation": false,
    "createdAt": "2025-01-14T10:30:00",
    "updatedAt": "2025-01-14T10:30:00"
  }
}

Notes:

Only the analysis owner can access their records
Returns 404 if analysis doesn't exist or user doesn't have access
Includes complete transcript with speaker identification

Data Models

Analysis Record Structure

{
  "id": 456,
  "userId": 1,
  "kbDocumentId": 123,
  "audioFilename": "sales_call_2025_01_14.mp3",
  "callDuration": "00:15:30",
  "clientSentiment": "Positive",
  "productsMentioned": ["Product A", "Product B"],
  "pricingMentions": ["$5,000 per month", "20% discount"],
  "actionItems": [
    "Sales Rep will send proposal by Friday",
    "Client will review and respond by Tuesday"
  ],
  "conversationSummary": "Brief summary of the call...",
  "transcript": [
    {
      "speaker": "Sales_Rep",
      "text": "Dialogue from sales representative"
    },
    {
      "speaker": "Client",
      "text": "Dialogue from client"
    }
  ],
  "alreadyEvaluated": false,
  "evaluatedAt": "2025-01-14T10:30:00",
  "isReevaluation": false,
  "createdAt": "2025-01-14T10:30:00",
  "updatedAt": "2025-01-14T10:30:00"
}

Field Descriptions:

id: Unique identifier for the analysis record
userId: ID of the user who owns this analysis
kbDocumentId: Reference to knowledge base document (unique per user)
audioFilename: Original audio filename
callDuration: Duration in HH:MM:SS format
clientSentiment: Client's sentiment - "Positive", "Negative", or "Neutral"
productsMentioned: Array of products/services discussed
pricingMentions: Array of pricing information mentioned
actionItems: Array of complete, actionable follow-up items
conversationSummary: 1-2 sentence summary of the call
transcript: Array of speaker turns with identified speakers
alreadyEvaluated: Whether this was a cached result
evaluatedAt: When the analysis was performed
isReevaluation: Whether this is an updated analysis
createdAt: Record creation timestamp
updatedAt: Last update timestamp

What Gets Analyzed

The Sales Call Agent automatically extracts:

Speaker Identification
- Separates dialogue between Sales Rep and Client
- Maintains conversation flow and context
Client Sentiment
- Analyzes overall client attitude: Positive, Negative, or Neutral
- Based on language, tone, and engagement level
Products & Pricing
- Identifies all products and services mentioned
- Captures pricing discussions, discounts, and negotiations
Action Items
- Extracts commitments and next steps from both parties
- Provides complete, actionable sentences with details
Conversation Summary
- Generates concise 1-2 sentence overview of the call

Action Items Extraction

The agent uses advanced prompting to extract complete, actionable sentences as action items:

✅ Correct Action Items:

"Sales Rep will send a detailed pricing proposal by end of day Friday"
"Client will review the product demo video and share feedback by next Tuesday"
"Sales Rep will schedule a follow-up call with the technical team next week"

❌ Incorrect Action Items (Avoided):

"Schedule demo" (too vague)
"Follow up" (no details)
"Send proposal" (missing who and when)

Sentiment Analysis

The agent analyzes client sentiment based on:

Tone and language used
Interest level expressed
Objections or concerns raised
Enthusiasm about products/pricing
Commitment to next steps

Sentiment Categories:

Positive: Client shows strong interest, enthusiasm, or commitment
Negative: Client expresses concerns, objections, or disinterest
Neutral: Client is exploratory, non-committal, or balanced

Error Responses

All endpoints may return the following error responses:

400 Bad Request:

{
  "detail": "document_id is required for security validation"
}

400 Bad Request (Short Transcript):

{
  "detail": "Transcript is too short or empty"
}

400 Bad Request (No Organization):

{
  "detail": "User must belong to an organization to use this agent"
}

403 Forbidden (Missing API Keys):

{
  "detail": "API keys not configured. Please configure in admin panel."
}

403 Forbidden (Document Access):

{
  "detail": "Access denied. Document 123 does not belong to user 1"
}

404 Not Found:

{
  "detail": "Audio file not found: sales_call_2025_01_14.mp3"
}

404 Not Found (Analysis):

{
  "detail": "Analysis not found or you don't have access to it"
}

500 Internal Server Error:

{
  "detail": {
    "error": "Transcription failed: [error message]",
    "error_type": "TranscriptionError",
    "message": "Unexpected error during analysis"
  }
}

Best Practices

Audio Quality Requirements

Format: MP3, WAV, M4A, or other common audio formats
Duration: 5-60 minutes optimal (longer calls supported)
Audio Quality: Clear speech from both parties
Recording: Use good quality recording equipment
Background Noise: Minimize for better transcription accuracy
Speaker Separation: Clear distinction between speakers helps

Recording Guidelines

Use a dedicated call recording solution
Ensure both parties are clearly audible
Avoid overlapping speech when possible
Record in a quiet environment
Test recording quality before important calls
Use headsets or quality microphones

Analysis Best Practices

First Analysis: Let the system cache results for faster future access
Re-evaluation: Use force_reprocess=true only when needed
Review Action Items: Verify extracted action items for accuracy
Sentiment Context: Consider sentiment in context of the full conversation
Follow-up: Use action items as a checklist for follow-up tasks

Integration Workflow

# 1. Upload audio to media directory (via your file upload system)
# Audio file: sales_call_2025_01_14.mp3

# 2. Analyze the sales call
curl -X POST "https://nextneural-api.superteams.ai/api/agents/sales_agent/analyze_call" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "file_name": "sales_call_2025_01_14.mp3",
    "document_id": 123
  }'

# 3. Review the analysis results
# - Check client sentiment
# - Review products and pricing discussed
# - Follow up on action items

# 4. Retrieve all analyses for reporting
curl -X GET "https://nextneural-api.superteams.ai/api/agents/sales_agent/my_analyses" \
  -H "Authorization: Bearer YOUR_TOKEN"

# 5. Get specific analysis for detailed review
curl -X GET "https://nextneural-api.superteams.ai/api/agents/sales_agent/analysis/456" \
  -H "Authorization: Bearer YOUR_TOKEN"

# 6. Re-analyze if needed (e.g., improved AI model)
curl -X POST "https://nextneural-api.superteams.ai/api/agents/sales_agent/analyze_call" \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "file_name": "sales_call_2025_01_14.mp3",
    "document_id": 123,
    "force_reprocess": true
  }'

Performance Considerations

Transcription Time:
- Short calls (< 10 min): 15-45 seconds
- Medium calls (10-30 min): 45-120 seconds
- Long calls (> 30 min): 2-5 minutes
Analysis Time: 5-15 seconds for AI analysis
Total Processing: Typically 30-90 seconds for a 15-minute call
Caching: Results are cached; subsequent requests return instantly
Concurrent Processing: Multiple calls can be analyzed simultaneously
Storage: Each analysis record is compact (typically < 100KB)

Security Features

User Isolation: All queries filtered by user_id
Document Ownership: Validates KB document belongs to user
Authentication: All endpoints require valid tokens with agent:sales scope
Organization Context: API keys scoped to organization
Unique Constraint: One analysis per document prevents duplicates
Access Control: Users can only access their own analyses

Use Cases

1. Sales Team Performance Review

Review sentiment trends across multiple calls
Identify successful conversation patterns
Track products mentioned and pricing discussions
Monitor action item completion rates

2. Follow-up Management

Extract action items automatically
Create follow-up tasks from analysis
Track commitments from both parties
Ensure nothing falls through the cracks

3. Pricing Strategy Analysis

Analyze pricing discussions across calls
Identify common objections or concerns
Track discount patterns and negotiations
Optimize pricing presentations

4. Product Feedback

Identify which products generate interest
Track feature requests and questions
Understand client priorities
Guide product development

5. Training and Coaching

Review call transcripts for training
Identify best practices from positive sentiment calls
Coach on objection handling
Improve sales techniques

Troubleshooting

Transcription Issues

Problem: Poor transcription quality

Cause: Background noise, unclear speech, low audio quality
Solution: Re-record with better audio quality, use noise cancellation

Problem: Transcription failed

Cause: Unsupported audio format, corrupted file, API issues
Solution: Convert to MP3/WAV, verify file integrity, check API key

Problem: Transcript too short

Cause: Audio file is empty or very short
Solution: Verify audio file has actual content, check recording

Analysis Issues

Problem: Incorrect sentiment

Cause: Complex conversation, mixed signals, sarcasm
Solution: Review full transcript, use force_reprocess for re-analysis

Problem: Missing action items

Cause: Action items not clearly stated in call
Solution: Review transcript, manually add missing items

Problem: Wrong speaker identification

Cause: Similar voices, overlapping speech
Solution: Review transcript, note for future calls

Performance Issues

Problem: Slow processing

Cause: Large audio file
Solution: Split large files, retry if timeout occurs

Problem: Analysis timeout

Cause: Very long transcript
Solution: Ensure audio is under 2 hours, retry if needed

Limitations

Language Support: Optimized for English; other languages may have reduced accuracy
Speaker Count: Best with 2 speakers (sales rep and client); multiple clients may reduce accuracy
Audio Length: Very long calls (> 2 hours) may have processing delays
One Analysis Per Document: Each document can only have one analysis record

Future Enhancements

Multi-language support with automatic detection
Advanced speaker diarization for multi-party calls
Emotion detection beyond sentiment (excitement, frustration, etc.)
Competitive intelligence extraction
Objection handling analysis
Talk-time ratio and conversation flow analysis
Integration with CRM systems for automatic follow-up creation
Real-time analysis during live calls
Custom analysis templates per industry
Trend analysis across multiple calls

Base URL​

Authentication​

How It Works​

Endpoints​

1. Health Check​

2. Analyze Sales Call​

3. Get My Analyses​

4. Get Specific Analysis​

Data Models​

Analysis Record Structure​

What Gets Analyzed​

Action Items Extraction​

Sentiment Analysis​

Error Responses​

Best Practices​

Audio Quality Requirements​

Recording Guidelines​

Analysis Best Practices​

Integration Workflow​

Performance Considerations​

Security Features​

Use Cases​

1. Sales Team Performance Review​

2. Follow-up Management​

3. Pricing Strategy Analysis​

4. Product Feedback​

5. Training and Coaching​

Troubleshooting​

Transcription Issues​

Analysis Issues​

Performance Issues​

Limitations​

Future Enhancements​

Base URL

Authentication

How It Works

Endpoints

1. Health Check

2. Analyze Sales Call

3. Get My Analyses

4. Get Specific Analysis

Data Models

Analysis Record Structure

What Gets Analyzed

Action Items Extraction

Sentiment Analysis

Error Responses

Best Practices

Audio Quality Requirements

Recording Guidelines

Analysis Best Practices

Integration Workflow

Performance Considerations

Security Features

Use Cases

1. Sales Team Performance Review

2. Follow-up Management

3. Pricing Strategy Analysis

4. Product Feedback

5. Training and Coaching

Troubleshooting

Transcription Issues

Analysis Issues

Performance Issues

Limitations

Future Enhancements