Rosalie: WhatsApp Order Automation, Replies Under 30 Seconds
Rosalie, an Indonesian maker of natural cheeses and snacks, was handling about 1,000 WhatsApp chats a week by hand. Biks built a WhatsApp AI agent that answers every customer in under 30 seconds, takes the order, confirms payment and books Gojek delivery, while the team approves exceptions. After months in daily use, admin time is down 2 hours a day and the closing rate has doubled.
- Customer chats answered
- 100%
- First reply, day and night
- <30 s
- Admin time saved each day
- 2 h
- Sales closing rate
- 2x
Figures from Rosalie, after months in daily use.

Before
- About 1,000 chats a week answered by hand
- Customers waited while the team juggled orders, payments and delivery
- Each order coordinated step by step
After
- Every chat answered in under 30 seconds
- Order, payment check and Gojek booking in one conversation
- The team approves exceptions and gets 2 hours a day back
WhatsApp was already Rosalie's busiest sales channel. Rosalie, an Indonesian maker of natural cheeses and snacks, was handling about 1,000 customer chats a week by hand, and its team spent hours each day answering questions, processing orders and arranging delivery.
Conversations were never the problem. The hard part was turning them into completed orders without making customers wait or taking control away from Rosalie's team.
How did Rosalie automate its WhatsApp orders?

Biks built a WhatsApp AI agent that gives each conversation a clear path from product question to confirmed order, payment check and Gojek delivery, while Rosalie's team keeps control of approvals and exceptions.
The workflow went live in three weeks, and Rosalie has run on it for months since. Customers kept buying through the channel they already used.
How does it work, step by step?
Every order follows the same path inside one WhatsApp conversation.
| Step | What happens |
|---|---|
| Customer asks | The agent replies in under 30 seconds and answers product questions. |
| Order | The agent takes the order and confirms the details with the customer. |
| Payment | The agent checks the payment; anything unusual goes to the team. |
| Delivery | Gojek delivery is booked from the same conversation. |
| Exceptions | Anything outside the rules goes to Rosalie's team to approve. |
What results did Rosalie get?
Figures from Rosalie, after months in daily use.
| Before | With Biks | |
|---|---|---|
| Customer chats | About 1,000 a week, answered by hand | 100% answered by the agent |
| First reply | When someone was free | Under 30 seconds, day and night |
| Admin time | Hours every day | 2 hours a day less |
| Sales closing rate | Baseline | 2x |
What changed for Rosalie's team?

- 100% of customer chats answered: Every customer gets a first reply in under 30 seconds, day and night.
- 2 hours of admin saved a day: The team reviews approvals and exceptions instead of processing every order.
- Closing rate doubled (2x): More conversations end in a completed order.
Instead of processing every order by hand, the team focuses on approvals, exceptions and customers who need a person.
See how Biks runs customer operations on WhatsApp, or read how another F&B brand takes orders after hours.
“With the new AI agent, we now own the entire order system end to end.”
Common questions
- How fast does Rosalie's WhatsApp AI agent reply?
- Under 30 seconds, day and night. Rosalie, an Indonesian maker of natural cheeses and snacks, answers 100% of its customer chats through a WhatsApp AI agent built by Biks, while the team approves exceptions.
- What does the WhatsApp agent do for Rosalie?
- It answers product questions, takes the order, confirms payment and books Gojek delivery in the same chat. Rosalie's team approves exceptions and has 2 hours of admin back each day.
- How long did Rosalie's WhatsApp automation take to go live?
- Three weeks from kickoff to live. Rosalie has used it daily for months since, and its sales closing rate has doubled.


