Shortening invoice flow from 12 to 2 days in a shipping company
We implemented an optical character recognition (OCR) system that automatically logs fuel and repair costs. Invoices are approved with one click on a phone.
A shipping company from Krakow was drowning in paper invoices for fuel and service. The accountant was losing 19 hours a month just re-typing data into Excel. We introduced a simple solution that does the same in seconds without errors.
The challenge
Drivers carried crumpled invoices in truck cabs for two weeks. Before the paper reached the office in Rybitwy, an average of 11 days passed. Attachments were often lost, which blocked VAT refunds and ruined relationships with fuel suppliers.
In February 2024, 83 documents were pending entry, and payment deadlines had long passed. The owner did not know exactly how much money the company spent in a given week because data flowed in with a huge delay.
Our approach
We started by visiting the transport base and talking to two drivers. We wanted to check why they didn't hand in invoices immediately. It turned out that looking for a post office on the route was a problem for them. We chose a mobile solution.
Our 3-person team configured an OCR system that reads Polish tax IDs and amounts from images. Instead of building a massive portal, we created a simple path: a phone photo goes straight to office verification. The entire work, from analysis to launch, took us 14 business days.
The solution
We launched a digital document flow connected to the company's current financial program. Now the driver takes a photo of the invoice right after refueling. The system automatically extracts the date, net amount, and VAT.
The shipping manager receives a notification on his phone and can approve the payment with one click while waiting for loading. There are no more binders circulating between desks. All data is visible in the panel 3 minutes after the photo is taken.
Results
Document cycle time dropped from 12 days to less than 2 business days. The accountant recovered 14 hours a month, which she now spends on monitoring deadlines instead of tedious data entry from crumpled sheets.
Timeline
-
January 8, 2024Interview with drivers and analysis of 47 typical cost invoices.
-
January 22, 2024OCR reader configuration and testing on dirty documents.
-
February 5, 2024Training for the office team on the approval panel.
-
February 19, 2024System launch for all 64 vehicles.
"I thought my drivers would grumble about a new app. It turned out they'd rather take a photo than guard papers in the cab for two weeks. Finally, I know how much I spend on fuel in real-time."