AI-SIM Final Seminar brought together research and practice to explore the role of AI in purchasing and supply management
The AI-SIM Final Seminar Gaining benefits from AI in purchasing and supply management was held in Tampere on 29 October 2026, bringing together researchers, company representatives and purchasing and supply management professionals to discuss how artificial intelligence is changing the field.
The seminar marked an important milestone for the AI-SIM project (AI-Driven Renewal of Business and Supplier Information Management), which started in the beginning of the year 2024 and ends by the end of November 2026. The project is a Business Finland Co-Innovation project, involving Tampere University, University of Turku, and industry partners AGCO Power, Integrata, and Sievo. The project has focused on investigating AI, especially Natural Language Processing (NLP) based solutions, for transforming supplier information management in procurement and supply chains.
The event attracted strong interest from practitioners and researchers. Around 80 people had registered for the seminar, reflecting the growing interest in understanding not only what AI technologies can do, but also how organizations can translate their potential into concrete business benefits.
From research results to practical AI applications
The afternoon opened with words from Aki Jääskeläinen, Professor at Tampere University, followed by presentations highlighting the results and experiences of the AI-SIM project partners. AGCO Power, Integrata, Sievo, Tampere University and the University of Turku presented different perspectives on applying AI in purchasing, supplier information management and related organizational activities.
The company partners demonstrated how the questions explored in the project translate into practical organizational settings. Jaakko Sundberg from AGCO Power presented its work on AI-supported procurement decision-making, including automatic component categorization and a similarity model combining component characteristics with price and supplier information. They were able to achieve promising accuracy in automatic categorization and identify significant savings opportunities. At the same time, one of the key lessons was that preparing and integrating the underlying data can require considerably more work than developing the AI algorithm itself.
Riku Honkanen from Integrata presented AI applications in HR and payroll services, including AI-assisted customer support and AI-based payslip analysis. The project also demonstrated how AI can move beyond supporting individual tasks towards coordinating more complete workflows across organizational and system boundaries. Saana Kivimaa from Sievo approached the challenge from the perspective of solution provider. Its work highlighted recurring challenges such as matching supplier records across different data sources, fragmented supplier information and difficulties in accessing critical external information. The project explored how AI can support the supplier data flow from harmonizing and enriching information to consuming it through AI assistants and, eventually, supporting or automating selected actions.
Research conducted within the project complemented these practical examples. Elviira Saarelma and Abul Khair Jyote from Tampere University briefly discussed different research findings related the project, including survey results on the level of AI implementation, the performance outcomes of AI usage, and the factors affecting the AI adoption in purchasing and supply management. Lastly, Laleh Davoodi from University of Turku presented research on the use of NLP and large language models (LLMs) in supply chain risk management. Her presentation highlighted how language models can be used to extract and classify risk signals from unstructured sources such as business news, while also emphasizing the importance of contextual understanding and human oversight in ensuring reliable results.
One theme connecting many of the presentations was the remarkable pace at which AI has developed during the AI-SIM project. When the project started in early 2024, the expectations were considerably different from where the AI technology development is today. As AI capabilities evolved, the original plans and areas of exploration also evolved across the project partners. What initially seemed like longer-term opportunities became increasingly realistic during the project, while new technological possibilities continuously emerged. At the same time, another recurring theme throughout the seminar was that the rapid development of AI does not remove the underlying challenges related to data. In many cases, making data accessible, reliable and usable for AI can require substantial effort before the potential of increasingly capable AI models can be fully realized.
NLP, language models and agentic AI looking towards the future
The seminar also looked towards the future through two keynote presentations. Filip Ginter, Professor from University of Turku, discussed the growing role of LLMs as critical infrastructure for research and development. His presentation explored how LLMs are developed and highlighted differences between closed, open-weight and fully open models. He also emphasized questions of transparency, reproducibility, privacy and technological sovereignty, arguing that fully open models could play an increasingly important role as organizations and societies become more dependent on AI.
Teemu Marttinen from Zeal Sourcing brought the discussion directly into the purchasing context with a market review of AI-assisted procurement. His keynote highlighted the shift from AI supporting individual procurement tasks towards automated workflows and, ultimately, ecosystems of AI agents working across different systems. Through examples covering areas such as spend analytics, sourcing, purchase-to-pay, and contract management, he illustrated how agentic AI could reshape both procurement technologies and the work of procurement professionals. A central message was that AI should not simply be added to existing tools but considered as part of a broader redesign of procurement processes and operating models.
Discussions continued beyond the presentations
An important objective of the final seminar was also to bring together researchers and practitioners and create opportunities for exchanging experiences. The programme therefore included time for networking and discussion alongside the research and company presentations. The positive feedback received after the event also reflected the success of the arrangements.
As the AI-SIM project approaches its conclusion, this also leaves an important message for the future. The technologies available for applying AI will continue to change, making experimentation and continuous learning increasingly important. At the same time, organizations need to invest in the less visible foundations of AI adoption, particularly their data, information management, system integration and organizational capabilities, to translate technological progress into sustainable business value.
