The Rise of AI in Procurement and Supply Management

Briefing Doc: The Rise of AI in Procurement and Supply Management

This briefing doc reviews three academic papers and one popular press article concerning the nascent use of AI in contract negotiation in purchasing and supply management (PSM). These sources provide a comprehensive overview of the current state of AI adoption in procurement, outlining both the opportunities and challenges that come with this technological advancement.

Main Themes

Automation vs. Augmentation: AI is poised to significantly impact PSM, both through automation of routine tasks and augmentation of human decision-making. This transformation has implications for organizational structures, buyer roles, and supplier relationships. (Spreitzenbarth et al., 2024)

  1. Maturity Stage and Potential: AI and machine learning (ML) in PSM are still in their early stages. While there's significant potential for value creation, further research and practical applications are needed to realize this potential. (Spreitzenbarth et al., 2024)

  2. Use Cases: The literature identifies various use cases for AI in PSM, categorized by strategic, tactical, and operational levels. These include procurement strategy development, strategic supplier management, supplier pre-qualification, cost analysis, negotiation support, automated negotiation, supplier selection, risk monitoring, ordering, and supplier evaluation. (Spreitzenbarth et al., 2024)

  3. Data Challenges: The effective use of AI relies heavily on data availability, quality, and integration. Procurement organizations often face challenges in obtaining sufficient and reliable data to train AI models effectively. (Spreitzenbarth et al., 2024)

  4. Autonomous Negotiation: A key area of interest is the development of autonomous negotiation bots capable of independently handling contract negotiations. This raises questions about the bot's ability to represent user interests, navigate complex scenarios, and ensure ethical and fair outcomes. (Dumiak, 2024; Baarslag et al., 2017)

Important Ideas & Facts

  • Procurement 4.0: The fourth industrial revolution is driving a new wave of digitalization in procurement, with AI and ML playing a central role. (Spreitzenbarth et al., 2024)

  • Redefining Procurement: AI has the potential to fundamentally change how procurement organizations are structured and governed, as well as the skills required of procurement professionals. (Spreitzenbarth et al., 2024)

  • Early Successes: Companies like Walmart, Maersk, and Vodafone are already using AI-powered bots for contract negotiation, primarily for indirect spend categories. (Van Hoek et al., 2023; Dumiak, 2024)

  • Scaling Up: The scalability of AI solutions is a key advantage. Pactum, for example, claims its chatbot can conduct 2,000 negotiations simultaneously. (Van Hoek et al., 2023)

  • Future Trajectory: As AI becomes more sophisticated, it's anticipated that procurement professionals will focus less on routine negotiations and more on strategic supplier relationships and exception handling. (Van Hoek et al., 2023)

  • Bot-to-Bot Negotiations: The prospect of AI agents negotiating with each other raises questions about potential unintended consequences and the need for careful oversight and research. (Dumiak, 2024; Baarslag et al., 2017)

Key Quotes

  • "Many talk about imitation of human minds, but the human mind is very complex. I associate AI with solving complex problems and automation that tries to solve problems like a human using a machine." (Spreitzenbarth et al., 2024)

  • "We often work with qualitatively bad data and not much data at all. Digitalization must be seen end-to-end; it is not just having an intranet and a laptop instead of a fax machine. Often data is collected several times without knowledge from the other silos and with very different approaches and partners." (Spreitzenbarth et al., 2024)

  • "In the future, I can imagine all sorts of agents in the real physical world negotiating with one another… Letting these bots run completely wild, I think, requires more research." (Dumiak, 2024; Baarslag et al., 2017)

Conclusion

AI is rapidly transforming the landscape of procurement and supply management, presenting both exciting opportunities and significant challenges. While early applications demonstrate the potential for increased efficiency and value creation, it's crucial to address data limitations, develop robust governance frameworks, and conduct further research to ensure the ethical and responsible deployment of AI in this domain.

Timeline of Events

1989:

  • Research on negotiation support systems, including the development of Negoplan, an expert system shell for negotiation, is published by Matwin, Szpakowicz, Koperczak, Kersten, and Michalowski.

1997:

  • Cook publishes research on case-based reasoning systems in purchasing, highlighting their applications and development.

1998:

  • Khoo, Tor, and Lee explore the potential of intelligent software agents on the World Wide Web for automating part procurement.

2001:

  • Jennings et al. publish a landmark paper discussing the prospects and challenges of automated negotiation, setting the stage for future research.

  • Lam, Hu, Ng, Skitmore, and Cheung develop a fuzzy neural network approach for contractor prequalification in construction projects in Hong Kong.

  • Choy, Lee, and Lo create an intelligent supplier management tool for benchmarking suppliers in outsource manufacturing at the Hong Kong subsidiary of Honeywell.

2002:

  • Kashiwagi and Byfield test the use of artificial intelligence systems in the State of Utah procurement process to minimize subjectivity bias in best-value procurement.

2003:

  • Choy, Lee, and Lo use a hybrid case-based reasoning and neural network approach to reduce cycle time for benchmarking potential suppliers.

  • Shore and Venkatachalam develop a fuzzy logic model for evaluating the information-sharing capabilities of supply chain partners.

2004:

  • Cheung, Lee, and Lo implement a prototype system for a multinational manufacturer using an agent-oriented and knowledge-based system as distributed artificial intelligence.

  • Degraeve, Labro, and Roodhooft conduct a case study on the total cost of ownership for purchasing services, focusing on airline selection at Alcatel Bell.

2008:

  • Hindriks and Tykhonov research opponent modeling in automated multi-issue negotiation using Bayesian learning.

  • Moghadam, Afsar, and Sohrabi develop a hybrid intelligent algorithm for inventory lot-sizing with supplier selection.

  • Guosheng and Guohong compare neural networks and support vector machines in supplier selection.

  • Caputo and Pelagagge explore parametric and neural methods for cost estimation of process vessels.

2009:

  • Sim, Guo, and Shi develop BLGAN, a system combining Bayesian learning and genetic algorithms to support negotiation with incomplete information.

  • Plebankiewicz proposes a contractor prequalification model using fuzzy sets.

  • Guo, Yuan, and Tian research supplier selection based on a hierarchical potential support vector machine.

  • Lee and Ou-Yang use a neural network approach to forecast supplier bid prices and estimate the possibility of a successful deal during supplier selection negotiations.

  • Faez, Ghodsypour, and O'Brien develop a model integrating fuzzy case-based reasoning and mathematical programming for vendor selection and order allocation.

2010:

  • The annual Automated Negotiating Agents Competition (ANAC) begins, providing a platform for researchers to test and compare their negotiation agents.

  • Kuo, Wang, and Tien integrate artificial neural networks and MADA methods for green supplier selection in a case study at a global electronics manufacturer.

2012:

  • Ferreira and Borenstein develop a fuzzy-Bayesian model for supplier selection.

  • Kang, Lee, and Yang present a fuzzy ANP model for supplier selection, applied to the IC packaging industry in Taiwan.

  • Wu and Barnes create a dynamic feedback model for partner selection in agile supply chains.

  • Vahdani, Iranmanesh, Mousavi, and Abdollahzade employ a locally linear neuro-fuzzy model for supplier selection in the cosmetics industry.

2014:

  • Jain, Singh, Yadav, and Mishra use data mining techniques to evaluate criteria at the pre-qualification stage of supplier selection.

  • Son, Leu, and Nhung propose a hybrid Bayesian fuzzy-game model to improve the negotiation effectiveness of construction material procurement.

  • Dimopoulos and Moraitis publish research on advancements in argumentation-based negotiation.

2015:

  • Baarslag et al. summarize the key developments and results of the Automated Negotiating Agents Competition from 2010 to 2015.

  • Nepal and Yadav present a Bayesian belief network framework for sourcing risk analysis during supplier selection.

  • Chou, Lin, Pham, and Shao develop optimized artificial intelligence models for predicting project award prices using construction project data from Taiwan.

2016:

  • Tata Consultancy Services conducts a survey indicating the adoption of emerging technologies for automating sourcing processes, including recommending potential suppliers.

  • Hosseini and Barker develop a Bayesian network model for resilience-based supplier selection.

2017:

  • Baarslag and Kaisers propose a decision model for eliciting user preferences in automated negotiation, highlighting the importance of the value of information.

  • Hofmann, Neukart, and Bäck discuss the applications of artificial intelligence and data science in the automotive industry.

  • Gunasekaran, Papadopoulos, Dubey, Wamba, Childe, Hazen, and Akter explore the use of big data and predictive analytics for supply chain and organizational performance.

2018:

  • Bienhaus and Haddud analyze the factors influencing the digitization of procurement and supply chains, emphasizing the emergence of Procurement 4.0.

  • Abolbashari, Hussain, and Saberi present a Bayesian network model for measuring and improving procurement performance in organizations.

  • Bodaghi, Jolai, and Rabbani develop an integrated weighted fuzzy multi-objective model for supplier selection and order scheduling in a supply chain.

  • Choi, Lee, and Irani utilize a big data-driven fuzzy cognitive map for prioritizing IT service procurement in the public sector in Russia.

2019:

  • Pactum, an Estonian startup, is founded with a focus on developing AI-powered bots for autonomous contract negotiation.

  • Walmart International pilots Pactum's technology in Canada for negotiating contracts with suppliers of goods not for resale.

  • Handfield, Jeong, and Choi explore the use of data analytics and cognitive analytics as emerging procurement technologies.

  • Spreitzenbarth and Stuckenschmidt demonstrate the potential of regression trees and Bayesian optimization for reducing uncertainty in supplier selection within a total cost ownership framework in a case study at a German automotive manufacturer.

  • Bäckstrand, Suurmond, van Raaij, and Chen discuss the use of purchasing process models for teaching purchasing and supply management.

2020:

  • Spreitzenbarth, Bode, and Stuckenschmidt conduct a literature review on the state of artificial intelligence in procurement, comparing it to its use in sales and marketing.

  • The UK government publishes guidelines for the regulation of AI and ML technologies in public procurement.

  • Deloitte publishes a study on the uptake of emerging technologies in public procurement.

2021:

  • Allal-Chérif, Simón-Moya, and Cuenca Ballester discuss the potential of artificial intelligence to redefine the purchasing function, calling it "intelligent purchasing."

  • Pactum secures US $20 million in funding from investors including Maersk, expanding its operations and client base.

  • Forbes reports on Pactum's success in automating negotiations for Walmart and other large companies, highlighting the potential cost savings.

2022:

  • Flechsig, Anslinger, and Lasch conduct a multiple-case study on the potential, barriers, and implementation of robotic process automation in purchasing and supply management.

  • Bodendorf, Merkl, and Franke explore the use of artificial neural networks for intelligent cost estimation in the manufacturing supply chain, conducting a comparative study at BMW.

  • Spreitzenbarth, Bode, and Stuckenschmidt analyze the differences in the adoption and implementation of artificial intelligence in procurement versus sales and marketing.

  • Pactum expands its operations to Mexico, Central America, and China, broadening its global reach.

  • Bloomberg reports on Pactum's continued growth and expansion into new categories, including transportation and some goods for resale.

2023:

  • Burger, Nitsche, and Arlinghaus discuss the concept of hybrid intelligence in procurement, questioning the perceived superiority of AI and advocating for a balanced approach.

  • Spreitzenbarth, a buyer at a German automotive software company, completes his doctorate focusing on autonomous contracting agents, observing a clear trend towards increased autonomy in negotiation processes.

2024:

  • Spreitzenbarth, Stuckenschmidt, and Bode publish research on designing an AI-powered purchasing requisition bundling generator, applying it to a case study at an automotive software organization.

  • Spreitzenbarth, Bode, and Stuckenschmidt publish a mixed-methods review of the state-of-the-art in literature and practice regarding artificial intelligence and machine learning in purchasing and supply management.

  • Michael Dumiak publishes an article in IEEE Spectrum, exploring the increasing use of autonomous negotiation bots in procurement and the potential for bot-to-bot interactions. He raises concerns about the unpredictable outcomes of fully autonomous negotiations and emphasizes the need for further research.

Previous
Previous

GenAI? Maybe not yet an inventor, but already much more than that!