How deflection is measured
Deflection Rate = Deflected inquiries / (Deflected inquiries + Agent-handled inquiries)
"Deflected" typically means a customer found their answer via FAQ, knowledge base, AI agent, or chatbot without escalating to a human.
The four deflection mechanisms
- Self-service knowledge base — searchable articles
- AI agent / chatbot — conversational answers grounded in your knowledge base
- Pre-emptive notifications — status-page updates, order confirmations, proactive messages that prevent the inquiry
- In-product guidance — tooltips, empty states, and contextual help that reduce confusion at the source
Why 30–50% is the modern target
In 2026, FAQ-heavy verticals (e-commerce, SaaS onboarding) routinely hit 50%+ deflection with AI. Complex verticals (BFSI, healthcare) typically land at 20–30%. Anything below 15% suggests under-investment in self-service.
Deflection vs deferred
Common trap: customers who "deflect" via a knowledge-base article but come back the next day to ask the same question weren't deflected — they were deferred. True deflection requires measuring re-contact rate.
Common use cases
- Reducing cost per ticket
- Scaling support without proportional headcount growth
- Improving first-touch resolution and CSAT
- Freeing agent capacity for complex work
Related terms
AI Agent
An AI agent in customer support is an autonomous system that decomposes a customer's goal into steps, picks the right tools (APIs, knowledge base, escalation), executes them, and handles multi-turn conversations — distinct from a scripted chatbot.
Customer Support Software
Customer support software is the system of record for handling inbound customer questions, complaints, and requests — typically combining a unified inbox, ticketing, knowledge base, automation, and analytics.
First Response Time (FRT)
First Response Time (FRT) is the time between a customer's first message and the first substantive reply from a support agent — one of the most-watched customer support metrics because it correlates strongly with CSAT.
MTTR (Mean Time to Resolution)
MTTR — Mean Time to Resolution — is the average elapsed time from when a support ticket is opened to when it's resolved, measured across a defined period.
See ticket deflection in Bublly
Bublly bundles unified inbox, AI agents, ticketing, and analytics — covering everything in this glossary in one platform.
Request a demo