Answering 'where is my order?' calls with AI: a UK guide
By Keith Hamilton · · 5 min read
If you sell online or by phone, "where is my order?" calls are ones your team takes again and again. Each call is short, but together they fill the day, and the answer sits in a system your colleague has to look up while the customer waits. This article explains how an AI agent answers these calls for you, what it needs to do the job properly, and where a person should still step in.
Why do "where is my order?" calls suit AI?
They follow the same pattern every time, and the answer already sits in a system. Order and delivery enquiries share four features:
- The conversation is predictable: who are you, which order, where is it
- The answer is held in your order platform and your courier's tracking
- A slow answer is frustrating, but a correct answer is quick to give once the data is in front of you
- Most customers want a date, not a conversation
That combination means an AI agent gives a specific, accurate answer without anyone on your team being involved. For retailers, order enquiries are the natural first job for AI customer service. For a wider look at what voice agents can and cannot do, see our guide to AI voice agents for customer calls.
What does AI need to answer order tracking calls?
It needs three things: a way to find the order, access to tracking, and your rules for late or missing orders. An AI agent that only knows your returns policy will not get far.
A way to find the order
The agent matches the caller to an order using an order number, or a combination such as the phone number they are calling from plus their postcode or email address. Decide which details are enough before you go live.
Access to tracking
The agent reads both your order system and your courier's tracking. Your order system knows when something was dispatched; your courier knows where it is now. With both, the agent says "it left our warehouse on Tuesday and the courier expects to deliver tomorrow" rather than "it has been dispatched".
Your rules for late or missing orders
Decide in advance what counts as late, when a customer is offered a replacement or refund, and who makes that decision. The agent explains your policy and gathers details. The decision itself follows clear rules or sits with a person.
How does an AI order tracking call run?
A well-designed AI order tracking call is short and follows six steps:
- The agent says it is an AI assistant and asks how it can help
- It confirms who the caller is and which order they mean
- It looks up the order and the latest courier update
- It gives a plain answer, such as an expected delivery date
- It offers to send a text with the tracking link
- It asks if there is anything else, and passes the caller to a person if needed
If the caller asks for a person at any point, they get one, or a callback if nobody is free.
What happens when a delivery has gone wrong?
The agent recognises late, lost or damaged parcels and hands the call to a person quickly. These are the calls where customers are most annoyed, and where a poor experience does the most harm.
When the agent transfers the call, your colleague sees the order, what the tracking says and what the customer has already been told. Nobody wants to explain a lost parcel twice.
How do you protect customers' details?
The agent confirms details the caller gives rather than volunteering them, and never reads out a full address or payment details. Order records contain names, addresses and phone numbers, so control what the agent says before it has confirmed it is speaking to the right person.
Call recordings and transcripts are personal data, so UK GDPR applies to how you keep and use them.
How do you reduce WISMO calls in the first place?
Send customers the answer before they need to ask. The cheapest "where is my order?" (WISMO) call is the one that never happens, so look at when customers ring:
- If many call the day after ordering, a clear dispatch email with a tracking link answers the question before they pick up the phone.
- If they call when a delivery date passes, a proactive text when the courier reports a delay heads off those calls.
The AI agent handles what is left.
How much staff time does AI save on order calls?
The saving equals the share of calls the agent fully handles, multiplied by the time each call takes. Here is the calculation for a team that takes 300 order enquiry calls a week, at 4 minutes each including the look-up:
- Total time: 300 calls × 4 minutes = 1,200 minutes, or 20 hours a week
- If the agent fully handles a third of the calls, it releases about 7 hours a week
- If it fully handles two thirds, it releases about 13 hours a week
It never handles all of them. Your team still deals with the late, lost and damaged orders, which are the calls that most need a person. To put a value on that time, see how to calculate the real cost of manual work.
How do you measure AI for delivery enquiries?
Record today's figures before you switch anything on, then track the same figures afterwards:
- The share of order calls answered rather than missed
- The share fully dealt with by the agent, without a transfer
- Customers who ring back about the same order within a few days
- The reasons calls are transferred to a person
- A weekly sample of calls reviewed for accuracy
Where Forwardcycle fits
Forwardcycle helps UK businesses connect AI agents to their order and courier systems and design the hand-over to people. We are running a proof of concept for a European fashion retailer in which an AI voice agent finds orders in the courier network and tells callers when to expect delivery. No results are claimed yet. Find out more about our customer experience services or take the free AI Readiness Assessment.