Detect frustration before churn
Tracks mood per turn and triggers empathetic responses or automatic human escalation when frustration repeats.
Real-time mood tracking (calm, confused, frustrated, angry, happy) adapts replies or escalates automatically—paired with post-call summaries and field extraction included in every conversation.
Tracks mood per turn and triggers empathetic responses or automatic human escalation when frustration repeats.
Generates an Arabic/English summary and evaluates whether the conversation met your resolution rules.
Extracts booking IDs, order numbers, or lead scores into clean JSON and pushes them to your webhook.
Post-conversation sentiment, summaries, and extraction are counted inside the same `conversation_id`.
Define per-agent pass/fail rules to audit policy compliance across every call.
Pull structured fields like order ID, intent category, and follow-up date into JSON automatically.
Human agents receive the mood flag and summary before picking up the handoff.
{
"conversation_id": "conv_91f2a8",
"sentiment": "satisfied",
"escalated": false,
"summary": "استفسر العميل عن حالة الطلب #4821 وتم تأكيد التوصيل اليوم الساعة 4.",
"extracted_data": {
"order_id": "4821",
"intent": "order_status",
"resolved": true
}
} No. Post-conversation sentiment, summaries, and structured extraction are included under the same conversation quota.
Yes, via REST polling (`/v1/conversations/{id}/analysis`) or signed HTTPS webhooks when a call ends.