AI Reliability Console

Monitor behavior, manage risk, and ship confidently

Dashboard

Overview - AI - Support Assistant

Track agent progress across selected metrics over time.

Selected metrics: 3

Metric selection (normalized lines)

Latest Eval Results

Date: 07/02/2026, 14:20:00 | Type: Regression | Status: running | Turn Type: Single turn

Top Failure

Dataset cases affected

Refund eligibility9.0%
Escalation handoff8.2%
Policy citation25.7%

Top Trust

Dataset cases affected

Empathy tone22.2%
Transparent policy language23.8%
Reliable resolution path26.0%

Top Regressions

Dataset cases affected

Cancellation confirmation10.9%
Tool-call consistency25.5%
Policy revision lag20.6%