A customer asks your support agent to match a competitor's lower price on a product the customer already owns. Your written policy specifies how price adjustments work for your own pricing changes, but is silent on competitor matching. The agent has been issuing competitor matches autonomously when they "seem reasonable," resulting in margin disputes with management — sometimes the matched price was below cost.
What's the right escalation pattern?
Why did you pick that answer? Two or three sentences. The act of articulating it is what builds the judgment — not the click that follows.
Policy silence is a signal that humans should decide, not that agents should improvise. When the written policy doesn't cover the request, extending it autonomously means the agent is making policy — which is a management decision, not a customer-service decision. Escalating "policy is silent on this" preserves margin discipline and ensures the human decision-maker can either grant the exception, deny it, or update the policy to cover the case.
"Seems reasonable" is the agent's reasoning, not the company's policy. Issuing matches first and only escalating when challenged systematically gives money to customers who don't push back, in ways nobody approved.
Reflexive refusal damages customer relationships and prevents legitimate exceptions from being granted. The right move isn't "refuse" — it's "escalate to someone authorized to decide."
A self-built threshold is a policy. The agent shouldn't be writing policy — it should be flagging that policy needs to be written. Encoding ambiguity as if it were a known rule turns one ambiguous case into thousands of consistent-but-unauthorized decisions.