Voice Agent Health Dashboard

Real-time AI Performance Monitoring by ResultantAI

Chris Mott | ResultantAI

⚠️ Prompt Drift Detected - Escalation Rate Spike

Issue: Agent began hallucinating appointment times on Nov 14, 2025 at 2:47 AM EST. 12 calls escalated to human operators in 3-hour window (normal: 2-3/day).

Root Cause: Prompt didn't explicitly handle timezone conversion when customer location differed from business hours reference.

Chris's Fix: Added explicit instruction layer: "Business hours are 9AM-5PM Eastern Time. Always convert customer's stated availability to ET before confirming. If customer timezone unclear, ask explicitly before booking." Result: Escalation rate returned to baseline within 4 hours.

Dashboard Controls

Agent Type

Date Range

Compliance Flags

Task Completion

94.2%

Successfully completed interactions

↑ 3.1% vs. last period

Escalation Rate

8.3%

Calls requiring human intervention

↑ 2.7% (drift detected)

Avg Response Time

1.2s

Time to first response

↓ 0.3s vs. last period

Data Accuracy

97.8%

CRM fields validated correctly

↑ 1.2% vs. last period

User Satisfaction

4.6/5

Post-call survey average

↑ 0.2 vs. last period

Compliance Score

99.1%

No policy violations detected

Stable

Metric Details

Performance Trends (Last 30 Days)

See This System in Your Business

This dashboard monitors the exact AI voice agent architecture I built for MCA Financial Services. I can audit your current Retell/Regal setup and deliver a custom performance report in 48 hours.

Schedule Your Free Audit →

```