​
Call Analysis Guide

Learn how to upload, analyze, and interpret call recordings using AuthentiVoice’s AI-powered analysis system.

​
Analysis Process Overview

1

Upload Audio

Upload call recordings in supported formats
2

Automatic Processing

AI transcription and fraud analysis begins
3

Review Results

Examine transcription, fraud indicators, and risk assessment
4

Take Action

Assign for review, export reports, or escalate

​
Uploading Audio Files

​
Supported Formats

MP3

Most common format

WAV

Uncompressed audio

M4A

Apple format

OGG

Open source format

​
Upload Methods

  • Web Interface
  • Bulk Import
  • API Upload
Drag and Drop
  1. Navigate to Dashboard or Analyses page
  2. Drag audio file(s) to the upload zone
  3. Files begin processing automatically
Click to Browse
  1. Click the upload button
  2. Select file(s) from your computer
  3. Confirm upload

​
File Requirements

  • Maximum file size: 500MB
  • Maximum duration: 2 hours
  • Minimum duration: 10 seconds
  • Sample rate: 8kHz or higher recommended
  • Bit rate: 64kbps or higher for best results

​
Understanding Analysis Results

​
Analysis Dashboard Layout

Call Analysis Detail View
The analysis page is divided into key sections:
  1. Header Information
    • Call ID and timestamp
    • Duration and file details
    • Overall risk score
    • Quick actions
  2. Transcription Panel
    • Full conversation transcript
    • Speaker identification
    • Timestamp markers
    • Search functionality
  3. Fraud Analysis Panel
    • Risk indicators detected
    • Confidence scores
    • Red flag highlights
    • AI reasoning
  4. Action Panel
    • Review assignment
    • Export options
    • Notes and tags
    • Status updates

​
Risk Score Interpretation

​
Key Fraud Indicators

  • Stress Patterns: Unusual voice stress levels
  • Speech Rate: Abnormally fast or slow speech
  • Hesitation: Excessive pauses or filler words
  • Consistency: Changes in accent or speech patterns
  • Urgency Tactics: Pressure to act immediately
  • Information Requests: Asking for sensitive data
  • Story Inconsistencies: Conflicting information
  • Scripted Responses: Robotic or rehearsed speech
  • Evasion: Avoiding direct questions
  • Manipulation: Emotional manipulation tactics
  • Authority Claims: False representation
  • Technical Jargon: Confusing technical language
  • Background Noise: Call center sounds
  • Multiple Speakers: Coaching in background
  • Technical Issues: Suspicious connection quality
  • Caller ID: Spoofed or suspicious numbers

​
Working with Transcriptions

​
Transcript Features

Pro tip: Use keyboard shortcuts for efficient navigation
  • Space: Play/pause audio
  • ←/→: Skip backward/forward 5 seconds
  • Ctrl+F: Search transcript
  • Ctrl+H: Highlight selection

​
Speaker Identification

The system automatically identifies different speakers:
  • Speaker 1: Usually the caller
  • Speaker 2: Usually the recipient
  • Unknown: Unidentified speakers
You can manually correct speaker labels by clicking the edit icon. Search functionality includes:
  • Keyword search: Find specific words or phrases
  • Regex support: Advanced pattern matching
  • Time range: Search within specific time segments
  • Speaker filter: Search only specific speakers

​
Analysis Actions

​
Review Assignment

1

Select Reviewer

Choose from available reviewers or auto-assign
2

Set Priority

High, Medium, or Low based on risk score
3

Add Notes

Include context or special instructions
4

Set Deadline

Optional deadline for review completion

​
Export Options

PDF Report

Comprehensive analysis report with all details

Audio + Transcript

Original audio with synchronized transcript

CSV Data

Structured data for further analysis

JSON Export

Complete analysis data in JSON format

​
Bulk Operations

Select multiple analyses to:
  • Assign to reviewers in batch
  • Export multiple reports
  • Update status or tags
  • Archive or delete

​
Best Practices

​
For Accurate Analysis

  1. Audio Quality
    • Use high-quality recordings when possible
    • Minimize background noise
    • Ensure clear speech from all parties
  2. Metadata
    • Include caller ID when available
    • Add context notes
    • Tag with relevant categories
  3. Timely Review
    • Review high-risk calls immediately
    • Set appropriate deadlines
    • Follow escalation procedures

​
Common Issues and Solutions

Poor Transcription Quality
  • Check audio quality and format
  • Ensure minimum volume levels
  • Consider re-uploading with better quality
Missing Fraud Indicators
  • Verify audio contains speech
  • Check language settings
  • Review manual indicators
Slow Processing
  • Large files take longer
  • Check system status
  • Consider splitting very long recordings

​
Advanced Features

​
Custom Analysis Rules

Create custom rules for automatic actions: 
// Example: Auto-escalate high-risk calls
if (riskScore > 80 && callDuration > 600) {
  assignTo: "supervisor",
  priority: "urgent",
  notification: "immediate"
}

​
Integration with External Systems

  • CRM Integration
  • Ticketing System
  • Compliance Platform
Link analyses to customer records

​
Next Steps

1

Upload Your First Call

Start with a test recording to familiarize yourself
2

Review Analysis Results

Understand all sections of the analysis
3

Configure Preferences

Set up your notification and export preferences
4

Create Workflows

Establish review and escalation procedures