Quick answer
A WhatsApp assistant answers a bakery's repeat questions from your own menu: price by size and flavour, whether a date is possible, delivery areas and fees, and your halal and allergen statement. Each order lands in a Google Sheet. Custom designs, tiered cakes and rush orders inside your cut-off go to the baker for a price.
- Paste your menu, lead times, delivery rules and allergen statement; the assistant answers from that.
- Custom designs and tiered cakes are never priced by the bot; the baker quotes them.
- Every order becomes a sheet row: date, size, flavour, message, delivery or pickup, name, number.
- It pairs to your existing shop number by QR; plans start at $19 a month.
At a glance
It answers
- Cake prices by size and flavour
- Lead time and cut-off for orders
- Delivery areas, fees and self-pickup
- Halal, allergens and ingredients
It hands to you
- Custom designs and photo references
- Same-day and rush orders
- Complaints and refunds
10 min
to set up
New chats
only, by default
Any
language
Bakeries and cake shops
answered by Replai, after hours
- How much is a 1 kg chocolate birthday cake? I need it Saturday21:14
- 1 kg chocolate fudge is $48, serves 8 to 10. Saturday is fine if you order by Thursday. Pick up in store or delivery within 10 km for $8. Which one?21:14
- Delivery. Can you write Happy Birthday Mia? [photo of a design]21:16
- 'Happy Birthday Mia' on the cake, no charge. The design in your photo is custom, so our baker will confirm the price and whether Saturday works. Address and number please?21:16
Illustrative exchange. Your bot answers from your own prices, hours and rules.


Cake orders are made in the evening, usually the week of the party, usually with a photo attached. "how much is a 1 kg chocolate cake?", "can it be ready by Saturday?", "can you make this?". The baker is elbow-deep in fondant, and the person answering WhatsApp is the same baker at 11pm.
Why bakery WhatsApp is a nightly job
Speed decides who bakes the cake. The MIT lead response study found that calling a web lead within five minutes rather than thirty makes contact 100 times more likely, and qualification 21 times more likely (InsideSales.com/MIT Lead Response Management Study, 2007). A cake enquiry at 11pm behaves the same way: the bakery that answers the price and the date first is the one that gets the order.
The questions repeat: price by size and flavour, whether a date is possible, delivery or pickup, a name on the cake, halal and allergens. The answers are on your menu, but customers do not read menus, they ask. And the customers who ask at 11pm and hear nothing until the next afternoon order from the bakery that answered.
What can a WhatsApp bot handle for a bakery?
Paste your menu with prices per size and flavour, your lead times and cut-offs, delivery areas and fees, pickup hours, and your halal and allergen statement. The assistant answers from that on your existing number, in the customer's language.
| Customer asks | Answered from |
|---|---|
| Price for a size and flavour | Your menu |
| Can it be ready by a date | Your lead-time rules and cut-offs |
| Delivery or pickup | Your areas, fees and pickup hours |
| Message on the cake | Your rules, free or charged |
| Halal, nuts, eggless options | Your statement |
| How to pay | Your terms and the QR in its notes |
An order becomes a row in your sheet: date, item, size, flavour, message, delivery or pickup, address, name and number. You bake from the sheet.
What should stay with the baker?
Custom designs, tiered cakes, edible prints from a photo: the assistant never quotes these. It gathers the photo, the date and the guest count and hands the thread to the baker for a price. Rush orders inside your cut-off go to a person too, because only you know whether the oven has space. Complaints about a delivered cake go straight to you with the photo attached.
An evening with it switched on
Tuesday night, five enquiries. Three are standard cakes: quoted, dated, delivery confirmed, in the sheet. One sends a photo of a two-tier unicorn cake for Saturday; the assistant says the design needs a baker's quote, asks for the guest count, and marks it for you. One asks whether the brownies contain nuts; the assistant answers from your statement. You read the sheet over breakfast and quote the unicorn.
Set up in 10 minutes
Scan the QR with the phone holding your shop number. Paste the menu, lead times, delivery rules and the halal and allergen statement. Set language to auto. Order a test cake from another phone and read what it says. Fix anything unclear and switch it on. New customers are answered from that moment; your existing chats are untouched.
Related
- WhatsApp for cafes and coffee shops
- WhatsApp for florists
- WhatsApp chatbot for cash-on-delivery orders
- A starter WhatsApp assistant for any small business
Want this on your number?
7-day free trial — no credit card, no API key.
Frequently asked questions
Can a WhatsApp bot take cake orders?
- It answers prices, sizes, flavours, lead times and delivery rules from your menu, collects the date, message on the cake, delivery address and name, and logs the order to a Google Sheet. Payment stays with your existing process, usually a QR the assistant sends from your notes.
What about custom cakes from a photo?
- The assistant can see the photo, but it is told never to price a custom design. It confirms what it can from your rules, such as flavours and lead time, and passes the photo and details to your baker for the quote.
Can it answer halal and allergen questions?
- From the statement you write: halal certification status, whether the kitchen handles nuts, eggs or dairy, and which items are eggless or vegan. It is instructed not to guess an ingredient list: anything your statement does not cover goes to a person, which is the right line for an allergy question.
Do I need a new number?
- No. It pairs to the number on your Instagram and packaging by QR, like WhatsApp Web. Plans start at $19 a month with the AI included, and the 7-day trial needs no card.
Sources
- About WhatsApp, WhatsApp, 2026-09-07
- Lead Response Management Study, InsideSales.com and MIT, 2007