At Prijedorčanka, where day-to-day operations involve a large volume of documents, procedures and technical instructions, artificial intelligence (AI) has found its place in one of the most important resources of any business system: knowledge. Through the AI FORWARD project, the company has acquired an AI solution based on Retrieval-Augmented Generation (RAG) technology, named “Prijedorčanka AI – Mozak” (“Brain”). The solution enables employees to find information in internal documentation more quickly and reliably.
Founded in 1972, Prijedorčanka specialises in fruit processing and the production of fruit distillates and concentrates. The company combines its traditional expertise with new technologies and ways of working. Even before joining AI FORWARD, employees had received internal training on artificial intelligence, and some repetitive processes had already been automated. The project was therefore a natural next step: an opportunity to develop a practical solution tailored to the company in collaboration with a team of experts and to introduce AI in a structured manner, guided by clear rules.
“We envisioned Mozak as the company’s central knowledge base, rather than just another search tool. Our quality management system documentation, procedures and technical instructions are now consolidated within a single system. Every piece of information can be traced back to its original document. This turns the knowledge we have built over decades into an institutional asset, rather than something that depends on individual employees,” the Prijedorčanka team explains.
From a Vast Collection of Documents to Accessible Knowledge
Before the solution was implemented, employees looking for specific information had to know which document contained it and search through it manually, or ask a colleague who “knew where everything was written”. This was particularly challenging when it came to equipment maintenance, as each machine had its own documentation. This way of working was time-consuming and relied on individual employees’ experience, while new employees needed more time to find their way around.
This challenge was identified as an area where AI could deliver tangible value. Through AI FORWARD, and in collaboration with IT specialists, a solution was developed that allows employees to search the quality management system documentation by asking questions in natural language, instead of manually reviewing large numbers of documents.
AI as a Support Tool for Employees
Based on a question submitted by the user, the system identifies relevant information in the internal documentation and presents it together with the precise source — the document, section and page — allowing every answer to be verified. If the documentation does not contain a reliable answer, the system clearly states this rather than guessing. The solution operates entirely locally, without an internet connection, ensuring that company data never leaves the organisation. A Procedure for the Management of AI Systems and Risks, together with accompanying forms, has also been introduced.
The particular value of this solution lies in the fact that AI does not replace employees’ existing knowledge and expertise, but makes them more accessible and easier to put to use. Instead of spending time repeatedly searching through documents, employees can devote more of their attention to tasks that require their experience, judgement and decision-making skills.
Practical Implementation Driven by a Real Business Need
Prijedorčanka is one of ten companies selected to implement AI solutions in practice through the AI FORWARD project. The process began with an analysis of business processes and an assessment of where the use of artificial intelligence could deliver the greatest value. This was followed by the development, customisation and testing of a solution tailored to the company’s needs, alongside employee training and technical support throughout implementation. The solution is currently in the pilot phase, during which its use in day-to-day operations is being monitored.
This approach demonstrates one of the key values of the AI FORWARD project: AI is not introduced as an end in itself, but as a response to a specific business challenge. In Prijedorčanka’s case, the focus is on making better use of existing knowledge and reducing the time needed to find information, while maintaining full control over company data.
From Traditional Business Operations to AI-Powered Support
Prijedorčanka’s experience demonstrates that the use of artificial intelligence is not limited to technology companies. On the contrary, businesses with a long tradition in traditional industries can also identify specific business processes in which AI can create additional value. The implementation has also shown how collaboration between small and medium-sized enterprises (SMEs) and local IT experts can result in a solution tailored to a company’s real needs.
“For us, Mozak is a foundation, not an end goal. During the pilot phase, the system covers quality management system documentation and equipment maintenance. In the long term, we envision it as a central knowledge base that we will gradually integrate with procurement, purchasing, production and sales. We are taking a structured approach, guided by priorities and supported by clear rules,” the Prijedorčanka team concludes.
The AI FORWARD project is implemented with financial support from the European Union and the Government of the Federal Republic of Germany (BMZ), within the EU initiative for private sector development — EU4PSD — which forms part of the broader GIZ SEDEP programme, additionally funded by the Swiss government.