Moving a conversation to someone better placed to resolve it — deliberately, with the context attached.
Escalation. Escalation in customer support is moving a conversation from its current handler to someone better placed to resolve it — someone with more expertise (functional escalation) or more authority (hierarchical escalation). An escalation transfers ownership deliberately, with context attached, instead of letting a hard conversation stall where it is.
The two flavours solve different problems. Functional escalation moves a conversation sideways, to the person who knows — a billing specialist, an engineer. Hierarchical escalation moves it up, to the person who may — someone empowered to approve the refund or make the exception. Triggers can be human judgement, a customer asking for a manager, an SLA timer approaching breach, or — increasingly — an AI handing off when its confidence runs out, the same principle behind an approve-before-send model where the human is the permanent escalation point.
What separates a working escalation from a hot potato is what travels with it: the thread history, the internal notes, what has already been tried. If the customer has to re-explain, ownership moved but the escalation failed.
Escalation is often read as failure, and that instinct quietly damages teams: agents sit on conversations they can’t resolve rather than admit they need help, and customers pay for the pride in days of silence. The honest framing is the opposite — a fast, clean escalation is a good outcome, and a slow wrong answer is not. Teams without a defined path get the worst version anyway: the shoulder-tap, the forwarded email with no context, the conversation that becomes nobody’s.
Escalations are also a diagnostic feed. The same topic escalating again and again is not a staffing problem; it is a missing document, a missing permission, or a policy nobody wrote down — something fixable once, upstream.
Escalation rate formula. Escalation rate = (escalated conversations ÷ total conversations) × 100 over a period. Read the rate as a trend, not against a universal target: a rising rate signals a growing knowledge or authority gap; a rate near zero on a hard product usually signals sitting, not skill.
Define the paths in advance — for each conversation type, who is the escalation target and what must be attached — and give front-line agents enough standing authority that routine exceptions never need the ladder at all. Then review the escalation log monthly for repeats: every recurring escalation is a candidate for a new article, a new macro, or a wider permission that retires that escalation permanently.
Give your customers faster answers and your team their evenings back.
Desktop and mobile. Free for solo.