Before
Carrier invoices arrive as PDFs. Someone checks each line against the agreed rate in a spreadsheet and emails the carrier when a charge looks wrong.
Logistics runs on documents and inboxes: bills of lading, invoices, packing lists, rate requests and status emails. AI reads them into your TMS and spreadsheets, so your team checks instead of typing.
You'll talk to Usama, who writes the code.
A typical pattern, not a client result.
Carrier invoices arrive as PDFs. Someone checks each line against the agreed rate in a spreadsheet and emails the carrier when a charge looks wrong.
The AI reads each invoice as it arrives, compares it with the agreed rate and lists the lines that differ. A person checks those lines and decides what to dispute.
Pricing, customs decisions, claims and anything you'd be held to. AI prepares the data. Your team signs it off.
Your TMS or freight software if it has an API or a database we can read, Outlook or Gmail, carrier and tracking APIs, Excel and Google Sheets, accounting tools. We'll confirm your exact setup on the call before you pay for anything.
One project we've worked on is a PDF bank-statement converter that reads messy PDFs into clean data. Ask on the call which past project is closest to your setup.
All source code goes into your GitHub. It runs in your accounts, and the AI model account is in your name. You approve each milestone before the next one starts. Nothing is held in our accounts, so there's no lock-in.
I'm Usama. I write the code on every project and I'm the person on your call. Samisoftz is me and a small team, so we take a few jobs at a time.
Usually, including scans. We test it on a sample of your own documents and show you how many it reads cleanly and which it flags.
If the TMS has an API or import, yes, after a person approves the data. If not, it can fill a sheet your team imports.
Yes. Load requests, carrier emails and rate confirmations follow the same pattern: AI reads and sorts them, your team prices and books.
Not on its own. It prepares the request for your team to price.
It flags anything that doesn't match, such as weights that don't add up or a missing reference, instead of guessing.
Bring three examples to a free 30-minute call.
You'll talk to Usama, who'd build it.
30 minutes · no slides · a written answer if it's not a fit