“Where is my order?” is the single most common support question in most product and service businesses, and it is one of the few where the correct answer already exists in a system nobody has connected to the inbox.
Key takeaways
Order and job status is unusual among support questions: it is not a matter of policy or judgement, it is a fact recorded in Shopify, Jobber, your ERP or the carrier’s system. The only reason a staff member is involved is that the customer cannot see it and the inbox is not connected to it.
Connect those two and the question answers itself, correctly, at whatever hour it is asked.
The rule that matters: the response reports what the system says. It does not estimate a delivery date the carrier has not committed to, and it does not reassure a customer that something has shipped because it usually would have by now.
An inferred delivery date that turns out wrong converts a routine enquiry into a complaint, and it does so at scale once automated. If the record is unclear, the honest answer is that a person is checking, followed by a person actually checking.
Most status enquiries cluster at predictable moments: after the order confirmation, around the expected dispatch, and when a promised window passes. A proactive update at each of those points removes a large share of the enquiries before they are sent.
Keep proactive updates informative and infrequent. A message when something changes is welcome; a daily reassurance that nothing has changed is noise, and customers opt out of noise.
Delivery across Canada involves genuinely long transit times to remote and northern communities, multiple carriers with different tracking behaviour, and customs steps on inbound orders. A status answer that ignores those reads as wrong even when it is technically accurate.
Handle them explicitly in the response logic, and use the customer’s own timezone and language preference where you hold it.
A late order is a status question. A late order for something needed for a specific date is a customer relations question, and the two look identical to a matching rule.
Route on sentiment and stated urgency as well as ticket type, so the customer who is genuinely stuck reaches a person rather than receiving a well-formatted tracking link.
A 30-minute call is enough to tell you whether AI pays for itself here.
