Artificial Intelligence Procurement Assistant: Enhancing Bid Evaluation

Abstract

AbstractIn modern business, maintaining competitiveness and efficiency necessitates the integration of state-of-the-art technology. This paper introduces the Artificial Intelligence Procurement Assistant (AIPA), an advanced system co-developed with Solita, a Finnish software company. AIPA leverages Large Language Models (LLMs) and sophisticated data analytics to enhance the assessment of procurement call bids and funding opportunities. The system incorporates LLM agents to enhance user interactions, from intelligent search execution to results evaluation. Rigorous usability testing and real-world evaluation, conducted in collaboration with our industry partner, validated AIPA’s intuitive interface, personalized search functionalities, and effective results filtering. The platform significantly streamlines the identification of optimal calls by synergizing LLMs with resources from the European Commission TED and other portals. Feedback from the company guided essential refinements, particularly in the performance of ChatGPT agents for tasks like translation and keyword extraction. Further contributing to its scalability and adaptability, AIPA has been made open-source, inviting community contributions for its ongoing refinement and enhancement. Future developments will focus on extensive case studies, iterative improvements through user feedback, and expanding data sources to further elevate its utility in streamlining and optimizing procurement processes.