CUSTOMER CASE | RETAIL
Document Understanding boosts invoice processing at Epta
Author
PAIJU KOIVULA
Share
Epta, a company specializing in refrigeration solutions and services, expanded its automation toolbox by implementing an AI-powered solution that streamlines processing purchase invoices. The automation, based on UiPath’s Document Understanding technology, enables the processing of invoices from over 200 suppliers. The three complementary automations together save approximately 75 hours per month. Epta plans to continue investing in automation, with a new automation project already in the pipeline.
Epta began collaborating with Sisua Digital in 2023 to improve operational efficiency and productivity. So far, four automations have been developed, and a fifth is currently being planned.
The solution leveraging UiPath’s Document Understanding technology was developed to support and enhance two existing invoice processing automations. Epta had two processes using the same invoice data.
Previously, manual invoice processing took several hours of work daily from multiple employees, as they compared ERP system data with information found on PDF invoices. The goal of automation was to save employees’ time for other tasks and increase processing accuracy.
How the Document Understanding automation works in practice
Epta has an extensive supplier network, so the AI model was designed with scalability in mind. The two processes involved handling different invoice data fields. To ensure flexibility for future automations, the model was also trained to capture additional data fields from invoices, making it easy to reuse this data later.
The invoice processing automation compares ERP system data with invoice details such as product quantities, prices, and product codes. If the data matches, the robot processes the invoices automatically from start to finish.
The purchase invoice processing automation compares and complements data from the invoice management system with information from PDF invoices and routes the invoices to the appropriate payment systems.
The Document Understanding solution supports these two automations by retrieving invoices from the system, interpreting their content using AI, and storing the extracted information in a database for use by the other two automations.
The solution scales efficiently
Document Understanding has significantly expanded the scope of the automations – now the automations can process around 200 suppliers, compared to only 15 previously. The AI model enables processing a much higher invoice volume, which is the most significant benefit of the solution.
“The automations process about 70% of all invoices end-to-end, while the remaining cases are predefined exceptions requiring manual review. The AI model’s success rate is around 80%. It identifies incomplete supplier information and speeds up invoice handling, saving considerable working time,” says Jari-Pekka Pesonen, Material Manager at Epta.
Together, these three automations support order processing and streamline the work of the finance and procurement departments. The three automations together save roughly 75 working hours per month, depending on invoice volume.
“The solution generates significant time savings and improves process quality as accuracy has increased. The operational purchasing team, in particular, benefits greatly from this automation,” Pesonen explains.
In addition, leveraging an AI model enables more efficient future development: when new suppliers are added, model training requires only a fraction of the effort compared to a traditional RPA implementation.
“We constantly onboard new suppliers, so it’s a big advantage that we can easily include them in the automation,” Pesonen notes.
“We are very satisfied with the Document Understanding solution: the automation has been reliable, with no errors, and the success rate has remained consistently high,” Pesonen adds.
Epta continues to actively expand its automation scope, with the next solution already in the pipeline
Epta holds regular robotics meetings to review process exceptions and improvement opportunities.
“With the robotics reporting, we receive a list of exceptions, allowing us to detect and resolve issues more quickly. This has sped up invoice processing and reduced errors,” says Pesonen.
The company intends to continue expanding its automation initiatives and is actively developing its collaboration with Sisua.
“We’re interested in expanding our internal robotics expertise through Sisua’s training – we’re considering training around five employees. Every department wants to leverage automation,” Pesonen says.
New automations are already being planned: for example, a robot will be implemented for the Order Desk team to assist with transfer pricing invoicing. Once completed, this automation will save about 48 working hours per month.
The collaboration between Epta and Sisua has been smooth, and Epta is very satisfied with the results.
“The Sisua team has been proactive in development and has quickly understood our needs. Their extensive experience with other companies has brought many valuable ideas to our joint meetings,” Pesonen concludes.
What was achieved
Benefits of the Document Understanding automation set
- Processes a significantly larger volume of invoices
- Saves approximately 75 working hours per month
- Frees up time for strategic work
- Improves process accuracy
- Easily adds new suppliers to the AI model
About the company
Epta
Epta Finland is Finland’s leading supplier of refrigeration solutions and services, with a factory in Porvoo and several sales and service locations across the country. Epta employs 328 people in Finland and is part of the global Epta Group. The company serves its customers by designing, installing, and maintaining refrigeration solutions throughout their entire lifecycle.
Book a free consultation with us
Want to know how intelligent automation could benefit your company? Tell us about your automation needs and let’s find the right solution for your company together.
