Deflection rate is a customer-service metric that measures the share of inquiries handled without a human agent — resolved or intercepted by self-service options such as FAQs, knowledge bases, and chatbots. It's widely used to gauge how much volume automation keeps off the support team's queue.
The important nuance is what deflection does and doesn't prove. A high deflection rate means a contact didn't reach an agent — but that isn't the same as the customer's problem being solved. Someone who gives up after a dead-end FAQ is counted as 'deflected' just like someone who genuinely self-served, so on its own the metric can reward avoidance rather than resolution.
That's why deflection rate is best read alongside outcome metrics. Resolution rate (the share of issues actually solved) and time to resolution show whether deflected contacts were resolved or merely abandoned. A rising deflection rate with flat or worsening resolution is a warning sign, not a win.
In the context of Zowie, the goal isn't deflection — it's automated resolution. Because Zowie's AI agent executes actions inside connected systems rather than redirecting customers to self-help, contacts get resolved in the conversation instead of bounced away from a queue. A deterministic Decision Engine governs what the agent is allowed to do, so resolution stays accurate and auditable rather than a guess.
Read honestly, then, deflection is a cost-side indicator while resolution is the outcome that matters to the customer. Teams evaluating AI should ask how much was genuinely resolved end to end — not just how much was kept away from agents.
In summary, deflection rate measures inquiries handled without a human, which makes it useful for capacity planning but misleading as a quality measure. Pair it with resolution rate and time to resolution, and favor platforms that resolve issues rather than simply deflect them.
Explore: Automated Resolution Rate, Time to Resolution, Flows