Transforming the Value of AI in Procurement
Over the years, businesses worldwide have been searching for effective ways to include AI in their operations. However, it’s unclear which implementation strategy is right for AI in corporations. Where would AI be the most valuable within business opportunities?
In a joint research collaboration with professors Ruomen Cui of Emory University and Shichen Zhang of Nankai University, Meng Li, associate professor and Bauer Fellow of the C.T. Bauer College of Business, explores the value of AI in procurement, specifically centering on a B2B (business to business) setting. Li’s presentation outlined the findings of his paper, “AI and Procurement,” published in the Manufacturing and Service Operations journal.
In the B2B market, chatbots have been used to assist procurement managers in routine procedures. However, according to Li, existing publications involving AI suggest that suppliers prefer to trust human judgment as opposed to chatbot recommendation–a phenomenon called algorithm aversion.
How could AI in chatbots be utilized in procurement with this phenomenon? Li focuses on AI’s two main functions–automation and smartness, which allow the software to augment human work through smart decisions by continuously learning, reasoning, deciding, and acting. Li details his field experiment, comparing suppliers’ wholesale price quotes across human and chatbot buyers under AI recommendation and no recommendation conditions.
By looking specifically at the impact of automation — i.e., the buyer uses a chat-bot to automatically retrieve prices instead of asking in person — and the impact of smartness — i.e., the buyer signals the usage of a smart AI algorithm in selecting the supplier—his results show that a smart recommendation system can reduce suppliers’ price quotations for chatbot buyers. With the incorporation of smartness in an automated system, AI can deliver the most value in procurement.
Li’s research confirms that there is a clear value of AI in procurement when both automation and smartness are utilized. However, in order to build on this potential, companies should initiate operations to improve their AI system’s smartness before moving onto higher levels of automation.