I asked 4 executives how they measure AI ROI. None started with AI tokens. (www.businessinsider.com)

At a recent summit in Paris, executives from leading organizations discussed measuring the return on investment (ROI) of artificial intelligence, revealing a significant shift away from focusing solely on token consumption—often referred to as "tokenmaxxing." Instead, they emphasized the importance of evaluating AI outcomes based on real business metrics, such as employee efficiency and customer satisfaction. Charles Holive from BNP Paribas CIB highlighted this approach, stating that true value comes from understanding how AI impacts productivity, rather than merely tracking the number of tokens used. This evolution in thinking reflects a broader trend within the AI/ML community, as companies like Amazon recently discontinued their internal AI-use leaderboards, which encouraged token-driven behavior. Executives from firms like Tata Consultancy Services and NTT DATA echoed the sentiment that measuring the effectiveness of AI tools should prioritize measurable business performance over sheer utilization metrics. While monitoring token usage remains important for managing costs, the emerging consensus suggests a need to align AI projects more closely with strategic goals and tangible value, ultimately ensuring that AI investments contribute meaningfully to organizational success.