Challenge #1 Current IT infrastructures Exis�ng corporate IT infrastructures are o�en disparate systems which are not set up for seamless integra�on of AI technologies. Many companies find that their hardware, so�ware, and data ecosystems are not ready for the heavy computa�onal demands of AI. Legacy systems and data silos, in par�cular, lack the agility and capacity required to support sophis�cated AI algorithms. This frustra�on with their current IT por�olio was especially noted by two industries at opposite ends of the AI sophis�ca�on spectrum: Consumer Packaged Goods, the industry most likely s�ll to be in the early phase of its AI explora�ons and least likely to see AI’s relevance to its current business model, and the Technology industry, which has some of the biggest demands for compute power to be able to develop and sell AI so�ware and services to all other industries. Challenge #2 Customer expecta�ons As AI solu�ons begin to interface more directly with customers, businesses are trying to ensure these technologies enhance rather than detract from the customer experience. Customer expecta�ons were ranked as the top challenge by the Banking, Financial Services & Insurance sectors and by the U�li�es industry — two very different but both highly regulated industries. Challenge #3 Current IT service providers Companies o�en depend on the exper�se and prebuilt AI technology solu�ons of external vendors as well as deployment and maintenance of AI systems, yet the quality and range of services offered can vary widely, affec�ng the pace and success of AI implementa�ons. (While not cited as the top challenge by any industry, current IT service providers was ranked second for both the Healthcare and the Life Sciences sectors.) 21 TCS AI for Business Study Key Findings Report