Answer stock questions without answering them
Most retail messages are the same handful of questions. The agent takes those, and your team gets the ones that are actually worth a person.
Is the linen shirt in stock in M?
Yes — M and L are in stock at the Mid Valley store. S sold out this morning.
Can I return it if it doesn’t fit?
14 days, unworn, with the receipt. Want me to hold one for you?
The problem
Half the day goes on "do you have this in blue", "what time do you close", "did my order ship" — and the messages that needed a real answer are buried underneath.
What it looks like
Customer asks about sizing, stock or store hours
The AI agent answers from your product pages and help articles
Anything it cannot answer goes to a named person, with the thread intact
Order and delivery updates go out as approved templates
Post-purchase follow-up runs on a flow, not on memory
The obvious objection
Will an AI give customers the wrong answer about stock?
It answers only from material you gave it — your site, your documents, your help articles. Anything outside that goes to a person rather than being guessed at.
Other industries
For restaurants and food businesses
Your customers already message you to order. Connected to WhatsMenu, the catalogue, the cart and the order status all live inside that same chat.
For service businesses
Quotes, bookings and follow-ups where the enquiry already arrived — with the basic questions asked before anyone picks it up.
Get your WhatsApp under control
Set up your number, invite your team, and let the AI agent take the repeat questions.