Every overnight
conversation, reviewed.
Every morning at 9 AM ET, Vectrant runs an AI review pass on every conversation from the previous overnight window (8 PM-9 AM ET). The review evaluates handling quality, missed escalations, hallucination risk, and customer satisfaction signals. A summary report is emailed to the team with the full transcripts attached.
AI auditing AI, every day.
Big Sandy's reference deployment hides the chat bubble during business hours and surfaces it overnight, when the live floor team is gone. The morning review is how the team learns what happened while they slept.
Per-conversation evaluation
Each overnight conversation gets a pass: was the customer's question answered? Was the answer accurate? Did the bot miss a moment that needed empathy? Did it hallucinate any data? Did the customer leave satisfied? Each dimension scored, with citations from the transcript.
Patterns surfaced
The morning report rolls up patterns: "47% of overnight conversations were about delivery timing — 12% had no good answer." "3 conversations triggered hallucination flags on the protection-plan price." "Customer satisfaction averaged 4.1 of 5; the rough patch was a refrigerator-availability question." Action items are explicit.
Quality canaries
Known-answer test queries (the AI quality canaries) run on the same schedule. If the canary set passes 100% the previous month and drops to 90% this morning, the team is alerted before a customer notices.
Overnight Reviews, running on your data.
30-minute walkthrough on your real catalog. No slideware, no demo data.