- Everything you need to know about invoice annotation (training).
Ocerra has its own data extraction engine that reads and extracts data from various invoice layouts. This engine is powered by the latest Artificial intelligence (AI) and Machine Learning (ML) technologies. It is a fascinating tool as it can be trained, so it learns and gets smarter every day.
What is invoice annotation?
Invoice annotation is the way to train the AI of the system to extra data from the invoice layout where the system did not extract it the first time.
This information needs to be on the invoice layout in order to be trained. If data is not present on the invoice layout, please check the data extraction section to see how to add data manually or set up default settings.
Our support takes care of invoice annotation when you start using the system. During the first two months after going live, we review documents that require training for you. This is done daily, after hours, so the next day you have a batch of new invoices ready to be processed from the previous day.
(!) Please note that we only use the annotation techniques to train Ocerra to pick up data that is clearly visible on the invoice layout. You might still need to add some data manually after annotation, for instance, if the invoice layout has an exception. You can check the most common examples here.
Ocerra is trained to read and extract the following information from the invoice:
- Invoice number
- Invoice date and due date
- Purchase Order number
- Supplier Name*
- Net or subtotal amount
- GST total amount
- The total or gross amount
- Tax Number
- Invoice table including line-item description, quantities and amounts on the line level
- Other cost, e.g. freight or discount
*The only exception that does not require the annotation technique. To ensure you have the correct supplier name on the invoice, the invoice layout needs to be matched with the supplier name from your system. Ocerra then remembers the match per invoice layout. See below for more information.
Learning to annotate
In Ocerra, we built-in a user-friendly annotation process that allows anyone to train the system.
Learning how to annotate is an essential skill when using an AI-powered system like Ocerra. It will help you to add information to the new invoice layout with a few clicks and make the system learn and remember it next time.
Annotation techniques to master
In Ocerra, you will find several annotation techniques depending on what data you would it to extract and remember next time.
Click on the link below to learn more about each process:
- Training a supplier name match on the invoice layout
- Training to pick up header-level information using value-based annotation
- Training to pick up the invoice table including lines and columns using table-based annotation
- Training to read only one line using single-line annotation (clean table feature)