Eight Global Companies Turn Targeted AI Into Measurable Returns

Corporate enthusiasm for generative AI has collided with a basic problem: many pilots lack benchmarks that connect model performance to profit. A review of eight multinational companies, including IKEA and Walmart, shows that returns are more likely when projects target narrow operating bottlenecks such as customer service, product moderation, procurement and customs clearance, with savings, revenue, speed and accuracy measured from the outset.

IKEA launched its Billie chatbot in 2021. From 2021 through 2023, it handled 3.2 million interactions, resolved about 47% of inquiries and saved 13 million euros. After 8,500 customer-service workers were retrained as remote design advisers, that channel generated 1.3 billion euros in fiscal 2022. Walmart said in July 2025 that its self-healing inventory system had saved more than $55 million, underscoring how tightly scoped AI can move from experimentation to measurable returns.