Glossary

The vocabulary used across the ResponZ product and documentation. One sentence each, written for a Customer Care manager or Head of Product.

Gap: A topic that drives support volume but has no dedicated article in your Help Center. A gap = an article to write.
Quality score Score, AI score: A 0–100 rating assigned to every reply (human or Fin Agent) on four criteria: relevance, accuracy, tone, completeness.
Fin Agent: Intercom's AI agent that auto-replies to a share of conversations. ResponZ scores it like a human agent so you know what it's really worth.
Cluster Topic cluster: A group of conversations covering the same topic, detected automatically. Used to measure volume by theme.
Sentiment: Customer emotion detected in the conversation: positive, neutral, negative. Helps spot frustration spikes.
CSAT Customer Satisfaction: The rating (1 to 5) your customers leave at the end of an Intercom conversation.
Containment: Share of conversations resolved by Fin Agent without human handover. The higher, the more load your AI absorbs.
FCR First Contact Resolution: Whether a conversation is resolved at first touch, with no reopen.
FRT First Response Time: Delay between the customer's message and your team's first reply. Reported as median and p90.
Churn risk: A composite score combining negative sentiment, reopens and trend to estimate how likely an account is to churn. A priority signal for Account Management.
Reliability: The correlation between the AI quality score and actual CSAT ratings. Above 0.7 means the AI score is a trustworthy proxy when CSAT is missing.
Window Analysis window: The historical depth of conversations ResponZ analyzes (all history, 30/60/90 days, or new-only). The single biggest lever on credit consumption.