An AI support agent for order questions on WhatsApp and calls
“Where is my order?” is the message a support team types the same reply to all day. The agent looks the order up in your store, shares the status and tracking link in the WhatsApp chat, and passes returns and damaged items to your team with photos attached.
Example workflow. Not client work.
An order question, start to finish
A customer messages on WhatsApp or calls about an order. The agent finds it by order number, or by the phone number the customer is using, and reads the status from your store platform and the courier’s tracking where both offer an API (integrations). Returns, exchanges, refunds, wrong items and damaged parcels become a ticket for your support team, as does any customer who asks for a person. The ticket or order note shows what the customer asked, what the agent said and any photos sent.
What it checks before sharing
Your scope lists what the agent may read back, such as order status, courier name and the tracking link. The scope sets which checks come first, for example matching the phone number to the order.
What it collects for a problem order
- order number and registered phone number
- the item in question
- issue type: not arrived, return, exchange, damaged, wrong item
- photos of the item and packaging, sent on WhatsApp
- pickup address for a return
- refund preference, for staff to approve
Marked delivered, but not received
The courier’s tracking says delivered and the customer says nothing arrived. The agent does not argue with either. It opens a “delivered, not received” ticket with the delivery scan time from tracking, asks whether a neighbour or guard may have taken the parcel, and leaves the courier follow-up to your team.
Where your team decides
Refunds, replacements and exceptions to your return policy wait for staff approval. The agent is set up not to issue refunds: it collects the details and photos, and your staff approve. It tells the customer their request is logged and when to expect a reply, using the timing you give it.
Pincode checks before buying
Shoppers can ask whether you deliver to their pincode, and the agent checks that against your courier’s serviceable list if you have one.
COD order confirmation calls before dispatch
For cash-on-delivery orders, we can design a confirmation call before dispatch. Reminder-style calls in general are covered on the follow-ups page. The agent reads back the items and the delivery address, and the call ends in one of four outcomes:
- Confirm. The order is marked ready for dispatch.
- Change the address. The agent takes the new address and pincode, reads both back to the customer, checks the pincode against the serviceable list, and holds the order for staff to update.
- Cancel. The agent notes the reason and marks the order for staff to cancel.
- No answer. The order stays in your confirmation queue, and your team decides what happens next.
DLT registration for outbound calls and messages is part of the build (follow-ups).
Stores we build too
We build online stores with Razorpay payments (Razorpay) and order management (websites). A support bot can sit on the same order data. Refund approvals and stock checks are automation work. How our bots answer from your own information is on the AI bots page.
Sample order check
Interactive simulation · sample data
Sample Threads is a made-up clothing store. The chat is about order SAMPLE-1042, placed by Kavya R. (fictional). Every status below comes from the sample order, not a courier.
Choose what happens
Pick one above. Nothing here is live: no call is placed, no message is sent and nothing is saved.
FAQ
Which store platforms can it read orders from?
Where your platform and courier offer an API; we name them in the scope.
Which language will the agent speak?
The one your callers use. Language is a project decision rather than a fixed list: tell us who the agent will talk to, and we test a model in that language before we commit to it in the scope. That includes languages spoken outside India.
What fills your support inbox?
Someone from our team replies within one business day.