What are conversation insights?

Conversation insights are AI-generated summaries and classifications that help your staff understand a thread quickly without reading every message from scratch.

They appear in the merchant dashboard as structured context—topics discussed, customer sentiment, products mentioned, and a recommended ticket action—so agents and managers can triage and decide next steps faster.

What insights include

For eligible conversations, insights can surface:

• Topics discussed — the main themes in the thread (shipping, sizing, defect, etc.).
• Customer sentiment — how the shopper’s tone reads across the exchange.
• Products mentioned — catalog items that drove the conversation.
• Recommended ticket action — a suggested outcome such as close, monitor, or escalate.

Together, these fields give a scannable briefing before an agent opens the full transcript or takes over from the AI.

Insights assist staff—they do not replace judgment

Insights are meant to assist staff, not replace human judgment. Your team should review the recommendation in the dashboard, accept it when it matches the situation, or override it when the conversation needs a different outcome.

For example, the AI might suggest closing after a routine tracking answer, but your agent may choose to monitor if the customer sounded uncertain. That override is expected—insights accelerate decisions; people remain accountable for final disposition.

Analytics and reporting over time

Over time, these structured insights also support analytics and reporting views. When topics, sentiment, and recommended actions are captured consistently, managers can spot patterns—rising defect themes, products with repeated confusion, or escalation rates—without manually tagging every ticket.

Use insights as a starting point: read the summary, confirm against the message history for high-stakes issues, then act in the inbox the same way you would for any other conversation.