Fully automated transformation removes thresholds for digitizing invoicing in Sweden

Purpose and goal:

The purpose of the project is to develop the technical platform to a higher degree of automation using innovative methods. The goal is to scale up and reach a wider market by employing a new generation deep learning model to automate the processing of manual invoices and convert them into the EU-standardized e-invoice format, Peppol BIS 3.0. By doing this, the technology can be applied extensively in the civil society and support the transition to e-invoices in society.

Expected results and effects:

With its upgraded service, eSynergetics can reach a broader market and assist more businesses. The service has the potential to save manual work in Sweden equivalent to millions of hours per year by significantly reducing barriers to digitalization and automation of invoice and payment flows. The savings are achieved by enabling a complete transition to e-invoices, which allows for minimizing systems, processes, and routines for handling manual invoices and optimizing operations based on 100% structured data.

Approach and implementation: The project is carried out in three phases, each step progressively ensuring final quality and delivery.

Phase 1: Improve training and testing data. The goal is to create high-quality data that enables training a model with high generalisability.

Phase 2: Enhance the model and system performance. With the improved training data, refine hte model to achieve higher generalizability and ensure quality.

Phase 3: Operational implementation. introduce the model into the production environment, optimizing it, monitoring its performance, and gradually transitioning the workflows.