AI coding – AI for automated coding

AI, or machine learning, which is the technical term, is Rillion’s latest innovation to streamline and improve the work of processing invoices. By activating this add-on service, Rillion One Invoice continuously learns more about how to post your invoices. Rillion One automatically fills in its best proposal for the coding, but at the same time gives users the option of correcting any mistakes.

No templates need to be created, no rules that control who approve a project or can approve spending for a cost centre need to be entered. All the coding is done in the background, completely automatically.

For a company with the right prerequisites, Rillion One can apply the correct posting proposal to more than 90% of supplier invoices.

Why use AI coding?

What AI coding can do for you:

• Minimize the time required for coding invoices
• Improve the quality of bookkeeping by reducing the number of errors in coding.
• The administrator no longer needs to create accounting rules. It means that the following tasks disappear; posting templates no longer need to be created and maintained, you no longer need to set up rules for the account to be used by a specific supplier, links between projects/cost centres and approvers no longer need to be created, etc.
• VAT management is easier for invoices from international suppliers. The system will automatically calculate the VAT and record it in the input and output VAT accounts.
• Skills transfer. Handovers to new staff or temporary staff get easier. AI coding means that Rillion One provides a model for how your invoices should be coded. So, there is less time needed to teach new employees what to do because the system knows how your invoices should be coded. 

Enable AI coding

AI coding is an add-on service and can be enabled in Rillion One under Add-on services.

Select the Enable AI coding (Automagic accounting) checkbox. Click on Save.

How it works

 1. Set up AI coding model

After you have been using Rillion One for a while, we can use all the invoices that are saved in your database and all your coding of these invoices to create an AI model. This model then becomes Rillion One´s way of storing the relationship between your invoices and your postings. The model could best be compared to the memories we gradually build up and use when we perform various tasks. The big difference is that a computer never makes any mistakes when it comes to remembering. It will always suggest as you use to do. Not as a human thinks they used to do.

2. Importing invoices and posting proposals

Each time an invoice is imported into Rillion One, it is compared with your posting model. The model looks for patterns in the invoice in the form of words and numbers to find other invoices with similar information. Then Rillion One proposes the posting that best matches the recently arrived invoice. 

3. Continuous learning

Each time you create a new posting, Rillion One will learn from it and include it as a potential posting proposal for future invoices.