Mixpanel for Voice AI
Map Your Voice AI Agent Call Journeys
See When And Why Your Voice AI Calls Are Failing
Call Path Analysis
Vizualize call flows and classify outcomes, KPIs and more!
- Map Caller Journeys
- Visualize common (and uncommon) conversation flows
Find the calls that go off course and figure out why! - Find calls that meet specific criteria. 'Show me all the calls that successfully booked an appointment.'
- Identify drop-off points Easily spot the call stage where frustrated callers hang up.
- Did we miss a key metric? Define your own!
Add your own custom metrics to track what matters most for your Voice AI Agent's performance.
'My key metric is the number of times a user repeats themselves. Show me calls with more than three caller-repeat events.'
Call Metrics
Take a look inside and see what's happening in your calls!
Mean Signal-to-Noise Ratio (dB)
% of Speech with Pitch > 300 Hz (%)
Mean Voice AI Latency (ms)
Mean Words Per Minute (WPM)
# Interruptions Per Call Per Minute (Int / Call * Min)
Improve Your AI Voice Assistant Quality
Login and get started with an API key!
Pricing plans for teams of all sizes
Flexible pricing plans designed to get you up and running in no time. Start monitoring your voice assistants today.
Professional
A plan that scales with your rapidly growing business.
$0.025/min
* per minute of call recording
Get started- Up to 50,000 call recording minutes per month
Enterprise
Dedicated support and infrastructure for your company.
Custom
Contact Founders
- Unlimited call recording minutes
Voice AI Quality Metrics
Speech-to-text errors? Your Voice AI agent misunderstands the human? It could be background noise.
Signal-to-Noise Ratio (SNR) in Voice AI
Signal-to-Noise Ratio (SNR) is a crucial metric in Voice AI. It measures the strength of the desired voice signal compared to background noise. A higher SNR indicates clearer audio, while a lower SNR suggests more interference.
Why SNR Matters:
- AI Performance: Higher SNR and better audio quality improve ASR accuracy and natural language processing.
- User Experience: Clearer audio means your Voice AI agent understands the user, improving call success rates.
- Adaptability: Monitor SNR to adjust audio signal processing in real-time for optimal performance.
SNR Ranges:
- Excellent: >20 dB
- Good: 10-20 dB
- Poor: <10 dB
By monitoring and optimizing SNR, Voice AI systems can deliver superior performance, ensuring clear and effective communication in various acoustic environments.
Background Noise
No Background Noise
Voice AI Evaluation
Easily find and fix issues with your Voice AI with these features.
Monitoring
Know what your Voice AI is doing with the real-time dashboard.
Monitor SNR, pitch, latency, WPM, and interruptions in real-time to ensure your Voice AI is performing well.
Mean Signal-to-Noise Ratio (dB)
Mean Words Per Minute (WPM)
Deep Dive
When your Voice AI call fails, find the issue right away.
Analyze detailed metrics for every conversation. Take a deep dive and resolve the issue quicker.
Conversation Flow Analysis
This timeline visualizes the back-and-forth between AI and Human, highlighting the flow of conversation and any interruptions.
Color Code: Blue represents AI speaking time, Green represents Human speaking time.
Interruptions: Red markers indicate points where one participant interrupted the other.
Duration: The length of each bar represents the duration of each speaking turn.
Overlap: Overlapping bars indicate simultaneous speech, often seen during interruptions.
Analyzing this timeline can provide insights into conversation dynamics, turn-taking patterns, and the frequency and nature of interruptions.
Evaluation
Make sure you didn’t break your Voice AI with your latest changes.
Whether it’s a new feature, a bug fix, or a third-party provider change, evaluate the impact on your Voice AI.
Daily Voice Metrics
These metrics represent the daily averages of key voice quality indicators. They provide insights into the overall performance and characteristics of the voice interactions.
Mean SNR
Mean Voice AI Latency
Mean WPM
Mean % Speech with Pitch > 300 Hz
SNR: Signal-to-Noise Ratio, higher is better.
Latency: Response time, lower is better.
WPM: Words Per Minute, measures speaking speed.
% Speech with pitch > 300 Hz: Percentage of time the voice pitch was above 300 Hz.
Know what your Voice AI is doing with the real-time dashboard.
Monitor SNR, pitch, latency, WPM, and interruptions in real-time to ensure your Voice AI is performing well.
When your Voice AI call fails, find the issue right away.
Analyze detailed metrics for every conversation. Take a deep dive and resolve the issue quicker.
Make sure you didn’t break your Voice AI with your latest changes.
Whether it’s a new feature, a bug fix, or a third-party provider change, evaluate the impact on your Voice AI.