Home
Line Item Classification
We built a machine learning pipeline to classify invoices into cost groups and types automatically for our customer .
Pilot Project Details
Our Pilot Project for Line Item Classification was for a German company operating in the Logistics industry.
Problem.
Thousands of monthly invoices
The company manages thousands of their clients' invoices. From truck related costs, transport, shipping, toll fees, port fees, to personnel related costs. These invoices get classified and analyzed for auditing and explorative purposes.
Accounting team is overwhelmed.
Apart from their ordinary workload, the accounting team has to carry the load of correctly placing these invoices in the correct groups, which is a tedious and very labour-intensive process. Dedicated positions are being filled for this particular purpose.
Most invoices are similar.
The issue the team is facing is mostly a quantitative one. There is simply too many invoices to be processed by employees. These invoices are very similar in nature, usually only differing in locations, times, or brand names, making them ideal for a machine learning approach.
Results.
85% accuracy on the
initial result
85% of invoices are classified to the correct class out of 39 total classes.
New Feature
for Clients
The new pipeline is being rolled out for over 100.000 users of the main app.
Major Cost
Savings
The automated classification saves many precious man hours
Interested?
Get in Touch
For inquiries regarding this project or to discuss potential collaborations, please contact us at projects@f02.ai.
Made with