AI for processing order history and sending SMS
We have Quick Book online as our main accounting program.
An automatic reminder system needs to be created to notify clients about upcoming orders.
Data about orders needs to be imported from Quick Book online:
- Clients
- Invoices
- Products (SKU)
Then the system needs to be trained to analyze for each client:
- order frequency;
- average interval between orders;
- seasonality of purchases;
- average check;
- product assortment;
- quantity of each SKU in orders;
- growth or decline dynamics of purchases.
Based on this data, the system should calculate:
- the date of the last order;
- the average reorder cycle;
- the projected date of the next order;
- the probability of repeat purchase (%).
Example:
Client: XYZ Restaurant
Average order cycle:
- MONIN Vanilla — 21 days
- MONIN Caramel — 28 days
- Angostura — 60 days
Forecast:
- next order: June 15
- probability: 87%
Automatic reminder module
The system should automatically contact clients before the projected reorder date.
Communication channel: SMS
Example message:
“Good day, John.
It seems that you may soon run out of MONIN Vanilla and Caramel syrups.
Would you like to repeat your previous order or make changes?”
or
“Good day, John
According to your purchase history, you may soon need to reorder MONIN Vanilla and Caramel.
Would you like to receive an invoice?”
“Can we add other products?
It seems you haven't ordered MONIN Lime Syrup 1L for a while.
We can help you create an order.”
AI response processing module
The system should understand natural responses from clients.
Examples:
- “Remind me next week”
- “Contact me in two weeks”
- “Not needed for now”
- “Send the price list”
- “Prepare the same order”
The system should automatically:
- postpone the reminder date;
- create a task for the manager;
- notify the responsible manager.
All communication history should be stored.
⸻
Client behavior learning module
The system should accumulate knowledge about each client:
- preferred communication method;
- convenient day for contact;
- convenient time for contact;
- typical purchase frequency;
- most frequently ordered products;
- products that have not been ordered for a long time.
The accuracy of forecasting should improve over time.
