Business customers send delivery orders as emails, spreadsheets and PDF forms. An AI reader extracts the addresses, contacts, parcels and time windows, matches the account and creates the dispatch job, and the customer gets a job number within minutes.
Typical results
≈15 hrs/wk
Dispatcher time no longer spent retyping orders
≈90%
Orders created without manual entry
≈75%
Fewer address and time window errors
Figures reflect typical results for this kind of automation.
The challenge
A courier's 40 business accounts each order in their own way: a one-line email, a spreadsheet of drops or a scanned order form. Dispatchers retype every order into the dispatch system, about 8 minutes each, while the phone rings. Typos in addresses or time windows send drivers to the wrong place, and customers wait to hear their order was received.
Before and after
Minutes to enter one order
−88%Order entry errors per month
−75%Minutes until the customer gets a job number
−92%The automation
Make watches the orders inbox around the clock. An AI reader built on OpenAI pulls pickup and drop-off addresses, contact phones, parcel count, weight, time windows and special notes from the email body or attachment. The sender is matched to an account in the CRM and the job is created in the dispatch system. The customer receives a confirmation with the job number, and orders with a missing address or unclear time are held for a dispatcher.
How it works
- 01
Order email arrives
Business customers send orders as free-text emails, spreadsheets or PDF order forms to the orders inbox.
- 02
AI reads the order
The AI reader extracts pickup and drop-off addresses, contact phones, parcel count, weight, time windows and special notes.
- 03
Account matched in the CRM
The sender is matched to a business account in the CRM, so pricing and billing details carry over to the job.
- 04
Job created in dispatch
A job is created in the dispatch system, and the customer receives a confirmation with the job number within minutes.
- 05
Unclear orders held
Orders with a missing address or an unclear time go to a dispatcher with the open question highlighted.
Starts from
- Order emails
- Spreadsheets and PDF forms
Delivers
- Dispatch job created
- Confirmation with job number
- Held orders for dispatcher
The impact
Order entry drops from about 8 minutes to 1 minute of review, saving dispatchers around 15 hours a week. Customers get a job number within minutes of sending an order, address and time window errors fall by about three quarters, and dispatchers focus on the few orders that need a question answered.
Questions about this automation
Can it read orders in any format?
It handles free-text emails, spreadsheets with one row per drop and PDF order forms, including scans. Account-specific details, such as a standard pickup address or a default time window, are stored in the CRM and filled in when an order leaves them out. Orders the AI cannot read with confidence are held, never created blindly.
What happens to orders with missing details?
If an address does not match a real location, a time window is unclear or a required field is empty, the order is held. The dispatcher sees the extracted fields with the open question highlighted, and can fix the field and release the job or send the customer a prepared question by email.
Which inbox, CRM and dispatch tools does it work with?
It reads the shared orders inbox in Gmail or Outlook, looks up accounts in the CRM and creates jobs in the dispatch system through its API. This is a multi-step build, usually rolled out in stages, starting with the business accounts that send the most orders.