What Is Support Ticket Deflection and How Does It Work?

Deflection happens when a customer gets their question answered without a human agent spending time on it. The three main deflection channels:

Knowledge base articles. Customer searches, finds an article, reads the answer. Done. No ticket created.

AI chatbot. Customer asks a question in the widget. AI responds from your articles. Customer confirms the answer helped. No ticket created.

In-context help. Tooltips, onboarding guides, or inline documentation that answer the question before the customer even thinks to ask.

Each deflected interaction saves $5-15 in agent time and provides an instant response to the customer — a win on both sides.

Deflection rate = (self-service resolutions) / (self-service resolutions + tickets created)

If 100 customers need help and 45 find their answer through self-service while 55 create tickets, your deflection rate is 45%.

What's good: - 30-40%: Decent — you have coverage for common questions - 40-60%: Good — your help center is working well - 60%+: Excellent — mature self-service operation

Track deflection rate over time. It should increase as you add articles and improve AI accuracy.

Add more articles. Every content gap that gets filled is a new category of questions the AI can answer. Write 2-3 articles per week based on your most common tickets.

Improve existing articles. If customers read an article but still create a ticket, the article isn't answering their question completely. Add the missing information.

Make help findable. Embed the widget on every page. Add links to your help center in your app's navigation. Put help links in transactional emails.

Train the AI. When the AI gives wrong answers, fix the underlying article. The AI improves automatically.

For a complete strategy, see how to reduce support tickets or learn what to write first. Start free.