Get a demo

Turn booking emails into clean order records

Read every booking email and extract origin, destination, cargo type, and weight into a clean order record. Loads land in your table without anyone typing.

Get a demo

The prompt

I want to automate order entry from inbound freight booking emails. Can you build me a flow that monitors my inbox, uses AI to extract order details like origin, destination, cargo type, and weight, and loads the structured records into a table?

Just copy and paste the prompt into a new Parabola flow to get started.

What Parabola builds

A workflow with six steps you can edit:

1. Monitor the inbox. The flow watches your shared booking inbox. New emails from customers, brokers, and shipper distribution lists land in the pipeline.

2. Classify the email. Is this a new booking, a revision to an existing booking, or something else entirely? Each type routes to the right handler.

3. Extract the order fields. Origin, destination, cargo type, weight, pieces, equipment type, pickup window, delivery window, customer reference. The AI step reads the body and any attached rate confirmations or load tenders.

4. Validate against the customer profile. Some shippers require specific equipment. Some lanes have known accessorials. The flow checks the booking against the customer's defaults and flags anything that doesn't match.

5. Write the order record. Push it to the TMS, the ERP, or a shared spreadsheet. The order ID gets returned in the reply so the customer has a confirmation.

6. Send the confirmation. Auto-reply from the same inbox with the order ID, the booked details, and any field that needs the customer to clarify before dispatch can move on it.

Why teams stop doing this manually

Booking emails come in every layout. Some customers send a structured rate confirmation. Some send free-text in the body with the details buried in a sentence. Some send a PDF tender and expect the operator to read it. Some forward a thread that has the booking three replies down. Every one of them lands in the same inbox and needs to land in the same TMS by the same deadline.

The manual version is one operator on the inbox treating it as a queue. Open the email. Read the body. Find the origin. Find the destination. Find the weight. Type it into the TMS. Click save. Reply with the order ID. Move to the next email. A team of two can do a few hundred a day on a good day. The volume scales every time a new customer onboards or a peak season hits.

The cost of being slow on order entry is direct. Bookings sit in the inbox while loads waiting to dispatch. Customer service has to call the customer to clarify a missing field that the email already had buried in it. Dispatch starts the day behind because the queue from the previous evening never cleared.

How it works

Step 1. Paste the prompt.

Open Parabola, paste the prompt in section 2, and let it ask follow-up questions about which inbox, your TMS, and the fields you want extracted on every booking.

Step 2. Connect your data.

Inbox for the bookings, TMS or ERP for the order write-back, customer directory for the profile validation. Optional rate-card files for the lane-level defaults.

Step 3. Run it continuously.

The flow watches the inbox, parses every booking, writes the order, and replies with the confirmation. The team works the exceptions, not the inbox.

Related prompts

FAQ

Does this work when the booking details are in a PDF attachment instead of the email body?

Yes. The AI extraction reads the body first and falls back to the attachment when the body doesn't carry the fields. PDF, Excel, and image-based tenders all flow through the same step.

What if a customer sends bookings in three different formats?

The classification step routes each one to the right extraction template. Add formats as you onboard customers. The flow learns each variant.

How does the flow handle revisions to an existing booking?

A dedupe step checks the customer reference and load number. New emails update the existing record. Conflicting fields get flagged so a human can confirm before the update writes.

Can the flow write to any TMS?

Yes. Most TMS vendors expose an API or accept a structured CSV import. The flow writes to whichever interface your TMS supports.

What if the booking is for a new customer the flow doesn't recognize?

It gets flagged and routed to a human for setup. The customer profile gets created once, and every subsequent booking from that customer flows automatically.
Bookings in. Orders in the TMS. No typing.
Paste the prompt, point it at your booking inbox, and let the order entry run continuously.
Start for free