Barré Technologies simplifies digital document management

Businesses in Hungary are faced with a shrinking talent pool and an aging workforce population. As a result, companies are looking to accelerate digitalization and automation as ways to facilitate future growth.

In many cases, tasks such as invoice processing are particularly labor intensive, with time-consuming manual checks against both internal and external data, offering prime opportunities for the new technologies.

“In document management and business process management, the workflows are pretty complex,” says János Prikk, Business Development Leader at Barré Technologies. “At Barré, we saw how end users sometimes struggled to find where they need to go to find the right data. We realized that an AI assistant that could interact with document and business process management applications could help users navigate processes more quickly and help existing teams to become far more productive.”

As experts in solutions such as IBM® FileNet® and IBM Cloud Pak® for Data, it was a natural choice for Barré to select IBM as the technology partner for this project. Barré chose the IBM watsonx® portfolio of products as the foundation for their AI assistant. IBM watsonx offers seamless integration with both IBM and third-party solutions, simplifying the development process for Barré’s engineers. IBM watsonx.ai® AI studio offers a large language model (LLM) with Central and Eastern Europe (CEE) language support, such as Mistral and Llama 3, that can be deployed on premises, offering enhanced security for commercially confidential information.

“Our target market sectors often handle highly sensitive data, frequently under strict government regulation, and it is important for them to keep data secure,” notes Prikk. “The on-premises model, enabled by the lightweight nature of the LLM, provides confidence that the Barré AI assistant offers the high security features they need.”

Working closely with the local IBM team, Barré designed and developed an AI assistant to support three principal use cases: invoice ingestion and processing, regulatory repository navigation, and business knowledge base navigation. The assistant responds to natural language queries by accessing enterprise application data in the background to retrieve the correct documentation and serving curated results rapidly and reliably.

Barré is working to introduce the AI assistant to their clients across Hungary. For example, a major hospital is currently testing the Barré AI assistant to process incoming invoices, post them to the in-house enterprise resource planning (ERP) system, validate its details online with the Hungarian tax authority, and approve and file them.

“From our own experience using the assistant within Barré Technologies as client zero, we’ve seen a 30% increase in efficiency when navigating enterprise applications,” says Prikk. “For a team of 50 employees, this translates to saving approximately 1,380 hours annually. At an average loaded cost of USD 65 per hour, that’s around USD 90,000 in productivity gains per year. Given the implementation and annual operating cost of about USD 100,000, we expect a full return on investment within 13 months.”

“We always start with a conversation to discover the business goals,” explains Prikk. “It is important to offer clients not only a technical solution, but a solution that incorporates all aspects of the business—finances, HR, and third-party partners.”

Prikk concludes: “Naturally, AI is a hot topic in Hungary, but most people are still not familiar with it. With IBM watsonx, we can add some magic to customers’ existing applications and offer a pragmatic solution that delivers rapid results, boosts productivity, and helps to solve the constraints of the workforce that will help Hungarian businesses grow.”

Based in Hungary, Barré Technologies (link resides outside of ibm.com) develops software solutions designed for enterprise content management and enterprise data asset management.