All glossary terms

Definition

Deflection Rate

Deflection rate is the percentage of incoming support tickets that are fully resolved by an AI or self-service system without ever reaching a human agent. A deflection rate of 70% means 7 out of every 10 tickets close automatically. It is the primary metric for measuring AI support ROI.

How it works

Deflection rate is calculated by dividing the number of tickets closed without agent involvement by the total number of tickets received, then multiplying by 100. A healthy deflection rate for an ecommerce store using AI support is 60–80%. SaaS companies with complex products typically target 40–60% as a starting point, rising as the AI knowledge base matures.

Deflection rate should always be paired with CSAT — a high deflection rate with a low CSAT score indicates the AI is closing tickets by discouraging customers from escalating, not by genuinely resolving their issues. True deflection means the customer's question was answered correctly on the first attempt.

When you'll encounter it

You'll see deflection rate in weekly support reports, AI vendor ROI discussions, and board-level customer experience reviews. It is the single most cited metric when comparing AI support tools because it directly translates to cost savings: each deflected ticket is a ticket your team didn't pay an agent to answer.

How SupportSyndicate handles this

SupportSyndicate reports deflection rate per channel, per time period, and broken down by topic cluster — so you can see whether order-status questions deflect at 90% while returns deflect at 55%. This granularity lets teams prioritize knowledge-base improvements for the lowest-deflecting categories rather than optimizing the average.

🍪 We use cookies

We use essential cookies for authentication and analytics cookies to improve our service. Read our Cookie Policy and Privacy Policy.