Success story #3 Real-world AI implementa�on: A masked interview with a large enterprise A revolution in retail An e-commerce firm integrates AI across sales forecas�ng, inventory management, infrastructure provisioning, customer service, and fraud detec�on Country: Australia Type of company: E-commerce retailer Role: Head of Analy�cs & Strategy Revenue: USD $20 billion to less than $50 billion Overview This seasonal business has peak demand around Christmas, Black Friday, and a few other promo�onal dates and was o�en challenged with shortages. As a result, there was overinvestment in inventory, forcing the company to carry it for a while, which was a loss in profitability. The various promo�ons, the compe��on, and ac�vi�es in the macroeconomic environment made demand forecas�ng challenging, The company wanted to use AI to do daily forecast demand, and it was able to do it down to the hour and minute with effec�ve website traffic management. The goal was to manage the en�re end-to-end supply chain inventory, store stocking, The explora�on of adop�ng AI technologies is extremely broad. We are trying to see how AI could be leveraged in each and every area of the organiza�on. Some areas include forecas�ng sales, predic�ng which products will be in demand, how much, during what season, and inventory planning. Another area is customer service, in which we're using automated chat to respond to customers without the staff involved. Then, we're using AI to scan SKUs in stores: a robot with AI that goes around on its own, plans its route and checks stock levels — so, essen�ally, inventory tracking and replenishment. Then, we are using AI for fraud detec�on. We look at all of that. Then, AI was used essen�ally to figure out, 'How much demand do we foresee on which day?' And we did it very granularly, actually down to the minute and hour. website traffic, and several load calcula�ons based on AI insights. AI was able to predict the number of orders to expect and gave the company lead �me to order the right products, and ensure the right products were in the right store during any period of demand. Another business challenge was the company website would crash during the biggest promo�ons, so it needed a way to es�mate the website load — essen�ally, determining how many hits are expected, based on the number of promo�ons, to allocate enough server bandwidth. AI was able to help the company with both product and cloud-based IT infrastructure alloca�on. TCS AI for Business Study Key Findings Report 18