Online retail
E-commerce
Pre-purchase questions answered from live product data, with order issues routed to a person.
These are illustrative scenarios
An illustrative scenario written to show how a workflow would run. It is not a description of an existing customer.
The situation
An online store gets the same pre-purchase questions all day — sizing, stock, delivery time, whether an item suits a particular use. They arrive on the website, on Instagram, and by email. Answers come eventually, but 'eventually' is a customer who has closed the tab. The team's actual time, meanwhile, goes on order problems, which is where it should go.
How the enquiry runs
Pre-purchase questions answered instantly from real data; post-purchase problems given to a human quickly.
- 1Customer
A visitor asks about sizing on a product page
Website chat, in English, on a specific product.
I'm 1.78m and usually wear a medium — will the large be too big in this jacket?
- 2System
The agent reads live product data
It answers from the actual size chart and current stock for that item, not from a generic paragraph about sizing.
- 3Agent
It answers and asks one useful follow-up
It gives the measurements, makes a recommendation based on the store's own fit guidance, and asks whether the customer wants the item in a specific colour — which is the question that most often decides the sale.
- 4System
The colour is out of stock
Rather than ending there, the agent offers to notify the customer when it returns and captures their contact with consent.
- 5System
The enquiry is recorded with intent
Product, size, colour, and stated intent to buy are written to the customer record, tagged as waiting on stock.
- 6System
Restocking triggers a message
When the item returns to stock, the customer is messaged once, on the channel they used, referencing the exact product.
- 7Customer
A different customer reports a delivery problem
An order placed six days ago has not arrived.
- 8Staff
Order problems go straight to a person
The agent retrieves the order status, confirms what it can, and escalates immediately with the order reference attached. Anything involving a refund, a complaint, or a delivery failure is a person's job by design.
What ends up recorded
Pre-purchase intent, which most stores never capture at all.
- Product and variant asked about
- Stated size, fit, or use case
- Whether stock was available at the time
- Consent to be notified on restock
- Channel and page the question came from
- Order reference, where the enquiry is post-purchase
Where the person comes in
The agent answers pre-purchase questions from live data and captures restock interest. Refunds, complaints, delivery failures, and anything touching payment go to a person immediately, with the order already looked up.
What this scenario depends on
- An e-commerce platform with API access to products, stock, and orders
- Accurate size charts and product data — the agent is only as right as the catalogue
- A rule for which enquiry types must escalate
- Consent capture for restock notifications, meeting your legal obligations
- Platform, messaging, and AI usage costs, billed to the store directly
Other scenarios
The same mechanics, in a different sector.
Beauty and wellness
An after-hours booking enquiry answered, qualified, and turned into a confirmed appointment.
Dental practice
A new-patient enquiry triaged and booked, with anything clinical routed to a qualified person.
Local services
A job enquiry qualified on location, scope, and timing before it reaches the person on site.