🎙️ AI Calls Analysis Tool

Powered by Claude AI for quality assurance

Analysis Settings

Team

Country

Select Agent

Time Period

From

To

Analyzing conversations...

Fetching call data from ElevenLabs

🎉

No Calls to Analyze!

All calls have been reviewed, or no calls match your criteria. Great work!

⚠️

Analysis Error

Analysis Results

ANALYZING...

0

Average Score across 0 calls

Avg Duration: 0s | LLM: 0ms | TTS: 0ms | ASR: 0ms

📝

System Prompt Issues

⚙️

Settings Issues

📞

Individual Call Breakdown

Agent System Prompt

Review and copy the updated system prompt with applied fixes

Select an agent first

Last Updated: Never

🔗 Open in ElevenLabs

Select an agent to load their system prompt...

📊

View Past Analyses

Coming soon with database integration

📈

Compare with Last Week

Coming soon with database integration

🧪 Test Your Fixes

Run test scenarios in ElevenLabs to validate your fixes worked

📋 Testing Instructions

Apply your fixes in ElevenLabs:
- System Prompt: Copy the updated prompt from the modal → Paste in ElevenLabs Agent Settings
- Settings: Apply each setting change manually using the paths provided
Test with these scenarios in ElevenLabs:

Scenario 1: Customer asks "Who is this?" or "What is Wecasa?" - Test if agent explains clearly without repetition

Scenario 2: Customer hesitates ("um...", "I'm not sure") - Test if agent handles uncertainty well

Scenario 3: Customer is busy - Test if agent offers to call back appropriately

Scenario 4: Test a full booking flow - Check confirmation clarity and closing
After each test call in ElevenLabs:
- Go to the conversation in ElevenLabs
- Copy the FULL transcript (all messages)
- Note the metrics: LLM latency, TTS latency, ASR latency
Paste everything below - Include transcripts + metrics from 2-3 test calls

Test Call Data:

Validation Results

✓ Updated System Prompt

Generated: