Artificial intelligence (AI) has presented itself as a disruptive innovation in many industries. Moving forward from automation, self-driving cars, algorithmic trading, and airplane simulation, AI has moved to a new application area, that of procurement. Cognitive procurement, a term coined by IBM, goes far beyond automation of manual and repetitive tasks. Today, the supply chain has to cope with increasing customer demand in terms of timely delivery, convenience, and customization. Procurement managers face challenges to manage such scenarios without making errors. This is where cognitive procurement comes into play. Cognitive computing facilitates supply chain managers to eliminate mundane tasks and assist them in decision-making in a complex environment.
Complications in the Procurement Process
Numerous factors complicates decision making in procurement organization. Although the goal is to curb spending and maximize product quality, various factors conflict with these two goals. For instance, using water routes for shipping would be economical, but would delay the delivery and thereby extend the time-to-market. Logistics, quality of raw material, payment terms, outsourcing, and offshoring all affect decision-making in a procurement organization. Additionally, external factors such as economic turmoil, laws and policy change, technological advancements, and environmental factors also complicate the decision-making process.
Applications of Cognitive Procurement
IBM’s Watson has approached this problem by accumulating massive amounts of data generated from transactions and financial records. In cases where numeric data is available, data analysts use machine learning to identify trends, correlations, and tendencies from historical data in order to build predictive models. In addition to predicting the future, cognitive procurement can also automate spend analysis, contract lifecycle management, and supplier management. The true power of cognitive procurement can be exhibited by Natural Language Processing (NLP) capabilities, where it can go through texts in the contracts and policies to help procurement teams simplify their daily operations and guide internal customers in purchasing. For instance, cognitive procurement can assist spot buys by evaluating the requisition, investigating the market to find the best value, and generate a new purchase order. Apart from this, cognitive procurement can be used to perform analysis to find the most efficient supplier and manage dynamic supply chain risks.
Challenges and Risks of Cognitive Procurement
Although the future looks promising with such immense capabilities, it would be challenging to implement such innovation rapidly. First of all, there will be an issue of accountability in case errors are caused by cognitive analytics. Secondly, not many organization have historical data in digital format, which would deem predictive models ineffective. People perceive such AI and machine learning systems as a black box, as they cannot see how they arrived at a particular decision. All they see is the inputs and the final decision, with everything in the middle being a grey area. The ultimate challenge in incorporating cognitive procurement would be the sheer amount of investment required to develop such AI.
For more information on cognitive procurement, IBM Watson, predictive analytics, machine learning, procurement system, and procurement analytics: