“Where is my order”: how to reply over WhatsApp in your online store

“Where is my order?” is the shipping-status question a customer asks after paying, and the industry calls it WISMO. It is rarely a complaint: it is someone who doesn't know what is happening with something they already paid for and wants reassurance.

That changes how to answer it. Treat it as a logistics problem and you reply with a tracking number. Treat it as an information problem and you reply with the status, the expected date and what happens next. According to Redo's WISMO guide, most of these contacts come from an information gap rather than a genuinely late shipment, and the share of support volume quoted online varies a lot between sources, so measure your own rate before setting targets.

Answering these messages by hand becomes a bottleneck because every reply needs the same steps: open the order, copy the tracking number, check the carrier's page and type it all back. That is several minutes of work for a two-line answer.

The real cost is accumulation, not any single message. During campaigns or peak season, status questions mix with sales, size exchanges and real complaints, and the agent ends up answering what arrived first rather than what matters most. A calm customer's simple question blocks the shopper who wants to buy right now.

That is why it helps to separate two kinds of message: those solved by reading data (status, date, tracking) and those that need judgment (lost parcel, wrong address, damaged item). The first kind can be automated; the second goes to a person with full context.

A good answer has four parts: it recognizes the order, states the current status, gives an estimated date and explains the next step. If one is missing, the customer writes again.

Recognize the order by product name or number, not a generic “your purchase”. State the status in customer language (“it left our warehouse yesterday and is in transit”), not the carrier's code. Give an honest date or range, even a wide one. And explain what's next: when the next update arrives, what to do if nobody is home, whom to write if three days pass without news.

An example for a clothing store: “Hi Camila, your order #1482 (2 items) left yesterday at 4 pm with the carrier and is in your city today. We expect it between tomorrow and Thursday. We'll message you as soon as it shows as delivered; if it hasn't arrived by Thursday, reply here and we'll check right away.” Three sentences cover everything.

When an order really is late, the right answer is to tell the customer first, acknowledge the delay plainly and offer a concrete action. Hiding the problem behind a tracking number is what annoys customers most.

Decide in advance which cases escalate: a set number of days with no movement, an address error, a parcel marked delivered that the customer never got, a damaged product. In those cases the AI should not improvise compensation or promise refunds; it should collect the data (order, photos if damaged, address) and hand the conversation to a person. We wrote a guide on doing that AI-to-human handoff without making the customer repeat everything.

A practical rule: if the answer depends on data that already exists in your system, let the machine answer; if it depends on a decision that costs money (reship, refund, a coupon), let a person decide.

Automating order-status replies works when the agent reads the real order data from your store and phrases it in your brand's voice, instead of repeating a fixed template. The difference shows in the first sentence.

With heysteff, Steff, the AI agent, connects to Shopify and reads orders and catalog, so it can answer on WhatsApp with the status of that specific customer's order. If you prefer, turn on copilot mode: the AI drafts the reply and a person approves it before it is sent, which is useful for the first weeks while you check how well it answers. For the full picture on Shopify stores, see the AI agent for Shopify.

Two recommendations: load your delivery times by region and your returns policy into the knowledge base so the AI doesn't invent dates, and set escalation keywords (“complaint”, “never arrived”, “damaged”) that route the chat to a person. Also make the proactive notice (“your order shipped”) and the reactive reply use the same language, so the customer doesn't feel they are talking to two different stores.

The cheapest way to reduce status questions is to set expectations well before and during the purchase. A clear date on the product page and a notice when the order ships prevent a good share of messages.

Three habits help: state on the product page how long dispatch takes and how long delivery takes (they are different and customers confuse them), send a message when the order ships with the tracking number and estimated date, and give customers a fast way to correct their address. Time-sensitive products such as food feel this even more; see our resource on order tracking for perishable food.

When customers ask anyway, that conversation is an opportunity: an attentive customer right after buying. A clear, fast reply usually improves how they see your store more than any campaign.

Can “where is my order” be automated without the customer noticing it's an AI? The reply can be automated, and the point is not to hide it but to make it useful: real status, a date and the next step. If the case gets complicated, it should pass to a person with the full context.

How fast should an order-status question be answered? Ideally immediately, because the answer lives in data that already exists. If a person handles it, set at least an internal maximum time and track first response time; our resource on first response time explains how to measure it.

What if the customer says the order shows as delivered but they never received it? This case should go to a person, not an automated reply. Ask for the address, a photo of the spot or details of anyone who might have received it, open a claim with the carrier and give the customer a concrete time frame for a response.

• AI agent for Shopify
• AI agent for ecommerce
• Copilot mode
• Answer questions before the purchase