Reopen rate

A ticket is closed; the customer returns on the same issue through the same or a different channel. That is a reopen, and the reopen rate is reopens divided by resolved conversations over a period. Helpdesks track explicit reopens (customer replies on a closed ticket) but miss recontacts through phone, a new chat or a different email, so an honest measure links conversations by customer and topic across channels within the window.

For AI support agents, reopen rate is the corrective to closed-ticket counts. An agent can close every conversation with a polite "glad I could help"; only reopen data reveals whether help occurred. It is therefore a component of the auto-resolution rate definition and a key guardrail alongside CSAT. It is segmented by intent and by handling path (AI, escalated, human) so that a specific failing intent can be pulled back to agent assist without switching off the whole system.

Reopen rate is not the same as repeat-contact rate for unrelated issues, and it is distinct from escalation rate, which counts handovers during the conversation rather than returns after it.

• Support manager: a rising reopen rate on AI-handled intents is the earliest sign that the agent is closing rather than solving.
• CFO: each reopen is a second contact, often through a costlier channel, that erases the saving from the first automated one.
• CTO: it is the metric that makes auto-resolution measurable; it needs cross-channel customer matching to be trustworthy.
• Product manager: reopens clustered on one intent often point at a product or policy problem the AI cannot fix.

Closed tickets are the easiest number to inflate, by humans and by AI alike. Reopen rate exists to expose that inflation by asking whether the customer had to come back. It prevents the specific failure of an AI support rollout that looks excellent on deflection and closure while quietly generating repeat contacts and churn. The trade-off is measurement effort: cross-channel matching and a defined window add complexity, and some reopens are legitimately new issues. Eazyware sets the window and matching rules in scope and reports reopen rate per intent from the first week of shadow mode.

• SaaS: detecting that AI answers on a new integration are being accepted but recontacted within days, indicating an incomplete article.
• E-commerce: linking a chat "resolved" return initiation to a phone call three days later about the same order.
• FinTech: tracking recontacts on dispute-status queries to catch cases where the AI gave a status the customer did not understand.
• Healthcare: matching booking-confirmation calls to subsequent calls about the same appointment.
• Logistics: identifying delivery reschedules that were confirmed but not executed in the dispatch system.

MetricA tracked outcome metric, not a skill. Eazyware defines the window and cross-channel matching during scope and reports it per intent for every Customer Service Agents deployment as a component of auto-resolution rate.

Eazyware service that covers it: AI Customer Service Agents. Starting prices are on the pricing page.