Deflection tracked by the specific route it happened through
Conversational AI · Ticket Automation
CSAT & Deflection Analytics
What never reached an agent, and how it landed
app.quipcom.ai
Illustrative UI
Illustrative QuipAI screen — Deflection funnel and CSAT for AI versus human
Ticket Deflection SearchAR
1,284
Conversations
64%+4%
Deflected
29%
Handed over
4.6 / 4.7
CSAT · AI vs human
Where it went
Answered by knowledge base58%
Self-service action (reschedule, resend QR)6%
Handed to agent · low confidence21%
Handed to agent · customer asked8%
Abandoned7%
Deflection · 8 weeks
In detail
What CSAT & Deflection Analytics does
Part of Ticket Automation in Chatbots, WhatsApp Bots & Knowledge AI. The desk work the AI does before an agent opens the ticket.
Accuracy measured over time rather than as a single snapshot
CSAT compared for AI-handled conversations against human-handled ones
Where it fits
The desk work done before an agent opens the ticket
CSAT & Deflection Analytics is 3 of 3 modules in this area. Each one works on the same record.
The Conversational AI workflow, end to end
CSAT & Deflection Analytics supports this flow alongside the other Conversational AI modules.
- 01
Ask
A customer asks on the website or WhatsApp.
- 02
Answer from knowledge
The bot answers from the knowledge base, with a citation and a confidence score.
- 03
Deflect or create a ticket
A satisfied answer ends there; an unresolved one becomes a categorised, SLA-bound ticket.
- 04
Agent replies with the assistant
The Smart Agent Assistant drafts the reply from the ticket, the knowledge base and the activity log.
- 05
CSAT & learning
Deflection, accuracy and CSAT feed back into where the AI is allowed to take more.
Who uses it
Who works with CSAT & Deflection Analytics
Support manager
Watches the live monitor, deflection and CSAT-vs-human numbers.
- Deflection counted by route
- CSAT for AI against human handling
- A live monitor of AI-held conversations
Related
More in Ticket Automation
app.quipcom.ai
Illustrative UI
Illustrative QuipAI screen — Category, priority, intent and summary assigned on arrival
AI Triage SearchAR
New · via email
Subject: Charged twice for the 14:45 sailing
Hi, I booked two seats on the 14:45 Harbour Cove sailing on Tuesday and my card statement shows two charges of ₹920. Please refund one. Booking HL-9012.
QuipAI triage On arrival
CategoryPayments › Duplicate charge
PriorityHigh
IntentRefund one charge
SentimentFrustrated, polite
Route toPayments desk · Dev Menon
Summary: duplicate ₹920 charge on HL-9012; customer wants one reversed. Gateway log shows two captures 4s apart.
AI Triage & Summaries
Category, priority, intent and a summary on arrival
app.quipcom.ai
Illustrative UI
Illustrative QuipAI screen — A bot conversation handed to an agent below the confidence threshold
Bot-to-Human Hand-over SearchAR
WhatsApp · +91 •••• 4021
We're a group of 40 — can we book the 11:00 with a private deck?
Group bookings above 25 need a quick check with our team. I'm connecting you to a person now — they have this conversation.
Confidence 54% — below your 70% threshold
Hi, this is Sana from Harbour Line. Yes — the upper deck seats 48. Which date?
Hand-over packetAgent has context
Reason Group size exceeds self-service limit
Summary 40 pax, 11:00 sailing, wants private deck
Customer Returning · 3 previous bookings
Suggested Upper-deck group rate, ask for date
Assigned Sana Ali · 12s to pick up
Bot-to-Human Hand-over
Below your confidence threshold, a person takes it
Other capability areas in Conversational AI
Bring your own FAQ or help pages.
We’ll load them into the knowledge base live and show the bot answering, citing and handing over.