Operations
Purchase orders in, production plan out
Singapore heritage food brand HarriAnns, which supplies hotels and runs eight cafes, replaced manual spreadsheet tallies of emailed hotel purchase orders with an AI platform that reads the orders, reconciles naming variations across 120+ product codes, and flags anomalies like a regular item suddenly missing. Owner Alan Tan says the tally dropped from a full day of admin work to under two hours of checking.
How it wentThe purchase-order tally dropped from a full day of admin work to under two hours of checking, per the owner.
Try it · based on the source's workflow
Read today's emailed purchase orders from [inbox or label]. Extract each order into a table: customer, product, quantity, delivery date. Match product names to our master product list, including these known name variations: [paste mapping]. Flag anything unusual: regular items missing, quantities far from the customer's norm, or new items you cannot match. Output the production tally in [format] and list only the exceptions for me to check. Never guess a product match: leave it flagged.
A starting template based on the workflow the source describes. Review it before you use it; you are responsible for what it does.