Chattbot för orderspårning: Svara på ”Var är min order?” dygnet runt

What Is an Order Tracking Chatbot?

An order tracking chatbot is an AI chatbot that can call a lookup tool. A plain FAQ bot can explain your shipping policy, for example "orders ship within two business days". It cannot tell a customer where their own parcel is, because that answer is different for every order and changes every few hours. To answer it, the chatbot must read live data from the system that holds it.

A good order tracking chatbot does five things:

• It understands any wording. "Has my package shipped?", "order 1042 status", and "still waiting for my shoes" are the same question.
• It asks only for what it needs. An order number, the email on the order, a postal code, or a tracking number.
• It reads live data. Status, last location, recent scan events, and the estimated delivery date.
• It explains the status in plain words. Then it tells the customer what happens next.
• It knows when to stop. Lost, damaged, and misdelivered parcels go to a person.

Order tracking is one part of e-commerce support. For the wider picture (product questions, returns, and recommendations), read our overview of how AI chatbots change e-commerce customer support.

Why Automate "Where Is My Order?" Questions

Support teams often call these questions WISMO, short for "where is my order". They arrive in waves: after a sale, before holidays, and every time a carrier has delays. They also arrive at night and on weekends, when nobody is on shift. Four things make them a good fit for automation:

• The answer already exists. An agent opens the store admin or the carrier's website and copies the status. A tool call does the same work in seconds.
• Speed matters more than depth. The customer wants one fact. If that fact takes three hours, they often send a second message to ask for an update.
• The questions are predictable. Most of them map to a small set of statuses: not shipped yet, in transit, out for delivery, delivered, or delayed.
• It frees your team for the hard cases. Damaged items, wrong addresses, and refund requests need judgment. Copying a status does not.

Where the Order Data Comes From

An order tracking chatbot is only as good as the system it reads. There are three common sources, and many stores use two of them together.

A common combination: the store tells the chatbot which order the customer means and which tracking number belongs to it, and the carrier tells it where the parcel is now. If your orders live in a system without a ready-made integration, you can connect it with your own API. Our guide to creating custom tools for your AI chatbot shows how.

Verify Before You Reveal: Privacy and Security

An order number is not proof of identity. Many stores use short, sequential order numbers such as #1001, #1002, and #1003. If your chatbot shows order details to anyone who types a number, a stranger can try numbers one by one and read other customers' orders. Security teams know this pattern from APIs as Broken Object Level Authorization (OWASP API Security Top 10, 2023): a system returns a record because the caller knows its ID, without checking that the caller may see it. A chatbot is one more door to the same data, so it needs the same check.

Use these rules:

• Ask for a second detail. The email on the order or the delivery postal code is something the buyer knows and a stranger usually does not.
• Do not confirm that an order exists when the check fails. "I could not find a shipment with those details" is safer than "That email does not match order 1042".
• Show the minimum. Status, estimated delivery, and a tracking link answer the question. The chat does not need the full name, street address, or phone number. This follows the data minimisation principle in Article 5(1)(c) of the GDPR (2016), which is a sensible default even outside the EU.
• Treat tracking numbers differently. Carriers show a public tracking page to anyone with the number. A carrier lookup that hides receiver details shows about the same as that page.
• Keep lookups read-only. Actions that change an order, such as a cancellation or a return, need a matching email or a person.

How to Design the Order Status Conversation

1. Ask for the right identifier, once

Tell the customer where to find it: "Your order number is in your confirmation email and starts with #." If they give a tracking number instead, use it. Do not send them through a form if they already typed the number in their first message.

2. Translate carrier language

Carrier statuses are written for logistics staff. "Shipment information received" usually means a label exists, not that the parcel has left your warehouse. A good reply says what the status means, what happens next, and links to the full tracking page for the details.

3. Use your own policies for context

Live data says where the parcel is. Your policies say whether that is normal. If an order is "not shipped" after one day and your processing time is two business days, the chatbot should say that this is on schedule. That only works if your shipping policy is in the chatbot's knowledge base.

4. Decide which cases go to a person

A clean transfer passes the order number and a short summary to the agent, so the customer does not repeat themselves. Our guide to AI chatbot human handoff covers how to define these moments.

How to Set Up an Order Tracking Chatbot with BubblaV

BubblaV is an AI chatbot that learns from your website and calls tools to read live data. For supported store platforms and carriers, order tracking needs no code. The DHL, PostNord, and nShift integrations are available on all plans. Handoff flows and custom tools need the Pro plan (see plans and pricing).

For details on each piece, see the integration pages for Shopify order tracking and nShift shipment tracking, the Flows automation builder, and BubblaV for e-commerce.

The Shopify integration guide lists every order tool, including returns and cancellations. The human handoff documentation explains how intent matching and tickets work.

Order Tracking Chatbot Launch Checklist

Common Mistakes to Avoid

Showing order details for an order number alone. This is the most serious mistake, because it can expose other customers' data. If you build your own lookup, require a second detail, such as the email or the postal code.

Answering with a link only. "Check the carrier's website" sends the customer away to do the work themselves. Give the status in the chat, then add the link for details.

Letting the bot improvise on exceptions. If the knowledge base has no lost-parcel policy, the AI may guess. A guess about a refund or a reshipment creates a promise your team must keep.

Ignoring stale store data. If your team ships orders but does not mark them as fulfilled in the store, the chatbot reports "not shipped". The chatbot reads your data; it cannot fix it.

The Bottom Line

An order tracking chatbot answers the most repetitive question in e-commerce support with live data, at any hour. Connect the systems that hold the answer, ask for a second detail before you show order data, explain statuses in plain words, and send delivery problems to a person with the context already attached. Then your team spends its time on the orders that actually need help.