1
Typed messages
Mixed Bahasa Malaysia, English and 中文 in one message.
Hundreds of WhatsApp customer groups. Three languages. Voice notes and handwritten purchase orders. We connected the order flow directly to AutoCount — without changing how customers order.
Orders processed each day
Missed orders
Returned to the team daily
Restaurants served
A quick view of the business, operating environment, and the workflow focus behind this deployment.
A live WhatsApp workflow connected to real order handling.
Discover the customer experience behind the automation.
Orders did not arrive through a form or portal. They arrived the way customers actually work — through WhatsApp messages across many customer groups, all day and often late into the night.
Mixed Bahasa Malaysia, English and 中文 in one message.
Orders spoken from the kitchen with no written record.
Handwritten or printed purchase orders sent as photos.
“Add two more boxes.” “Change delivery to Thursday.”
The customer had already placed the order. Everything after that was manual transcription.
Orders waited until someone was free to key them in.
Misread handwriting, wrong items and wrong quantities.
The team kept doing data entry late into the night.
More orders meant more manual workload.
Keep the customer workflow. Automate the order entry behind the scenes.
We built an AI order agent that connects customer conversations directly to AutoCount.
Text, voice notes, photos and PDFs.
Matches customers, items, UOM and previous pricing from AutoCount.
Writes directly into AutoCount through its document interface.
Follow-up changes amend the same order.
Anything unclear goes to the internal team for review.
The agent only does the work it is approved to do.
Approved data only.
Runs only at the agreed time.
Acts after the conversation settles.
One group, one account.
Confirmation only.
Unclear cases go to the team.
We’ll look at your actual order flow and tell you whether this approach makes sense for your operation.