MANThLE

GlossaryUpdated 3 min read

Call deflection

Call deflection is the handling of a customer contact that would have reached a phone agent in another channel instead, usually self-service, chat or a proactive message. It counts as a success only when the customer's issue is resolved there. A customer who gives up and calls back later was obstructed, and a good deflection rate excludes them.

Two contact centers can report the same deflection rate and run very different operations. In one, customers get their answer in the app and never think about calling. In the other, customers fight through a menu, give up, and call the next day angrier than before. The number hides the difference unless it is measured on resolution.

How is call deflection rate measured?

The common formula divides contacts handled outside the phone channel by total contacts. That version counts a session as a success the moment it starts. A more honest version counts a contact as deflected only when two things are true: the issue was resolved in that channel, and the same customer did not come back about the same issue within a set window, such as seven days, in any channel.

Measure it by issue type. Password resets and payment confirmations deflect easily. Questions about a specific charge, a missed delivery or a changed schedule usually do not, and a blended rate hides where the calls are coming from.

Published benchmark ranges for deflection rate exist, mostly from software vendors, but they mix deflection, containment and resolution, and rarely say which contacts were eligible. They do not compare like with like. Your own baseline, by issue type, is the comparison that means something.

Why does deflection obstruct instead of resolve?

Gartner, as reported by CX Today in August 2024, surveyed 5,728 customers and found that only 14 percent of customer service issues were fully resolved in self-service. Even for issues customers called "very simple," only 36 percent were handled there. Of customers who started in self-service, 45 percent said the company did not understand what they were trying to do, and 43 percent could not find content relevant to their issue.

Those numbers describe self-service built as a filter in front of the phone queue. It answers generic questions with generic content, and the customer's actual question is specific.

Why do billing inquiries turn into calls?

Billing questions are about one account. "Why is my bill higher this month?" has an answer, but it lives in the plan history, a promotion end date, a proration, a payment that has not posted or a one-time fee. An FAQ page cannot see any of that. The customer calls because only an agent can look at the account.

The same pattern sits behind most expensive call types: the question needs account context and sometimes an action, such as a credit or a payment arrangement. When nobody answers it, the question can harden into a billing dispute.

Where does AI change call deflection?

AI can answer the specific question with the customer's own data, inside the authenticated channel. It can explain the line item, show the proration, and apply a credit when policy allows it. It can also act before the contact exists, by explaining an unusual bill before it arrives.

The calls that remain should be the ones that need a person: disputes, hardship cases, complaints with regulatory weight. Those reach an agent with the account history and the AI's reading already attached, so the agent starts at the decision and skips the search. The full design for billing is on the billing inquiry automation page.

A deflection rate set as a target, with no resolution check, will be met by making the phone number harder to find. Track repeat contacts and complaint volume next to it, and read the escalations the way management by exception intends: as the record of what the process still gets wrong.

Questions

Call deflection is resolving a customer's issue in a channel other than a live phone call, such as self-service, chat or a proactive notice, before the customer needs to call.

Divide the contacts fully resolved outside the phone channel, with no repeat contact on the same issue within a set window, by all contacts that were eligible to be handled that way.

A good rate depends on the contact mix, and published benchmarks mix different definitions, so compare against your own baseline by issue type and track repeat contacts alongside it.

Customers get frustrated when self-service blocks the way to a person without answering the question, which pushes the same issue back into the phone queue later.

AI improves deflection by answering with the customer's own account data, taking permitted actions such as applying a credit, and passing harder cases to an agent with the context attached.

Sources

  1. CX Today, Only 1 in 7 Customer Service Queries Resolved with Self-Service, Gartner Study Finds (August 2024)
SolutionsBilling inquiry and invoice dispute automationBilling inquiry automation and invoice dispute management: AI explains the charge from the account, credits within policy and sends real disputes to people.GlossaryBilling disputeA billing dispute is a customer's formal challenge to a charge. How it differs from an inquiry or chargeback, the four stages operators run, and where AI fits.GlossaryManagement by exceptionManagement by exception means routine work runs to standard and managers see only the deviations. Where Taylor named it, how it works and what AI changes.IndustriesAI in telecom operationsTelecom AI for operations, not networks: order fallout, billing inquiries, disputes, collections and service changes, run by AI with people on the exceptions.IndustriesAI in public transit operationsAI in public transit operations: disruptions, open work, paratransit and customer contact. Exception-heavy, safety-bound and union-aware, people in the loop.