Report 08 Apr 2026

Expert View: Maximizing AI ROI Through Smarter Model Usage

Employees are increasingly using foundation models such as GPT, Claude, Gemini, and Perplexity in their daily work. As adoption grows, AI usage will directly affect the bottom line because prompts incur real costs, billed in tokens. Token consumption is rising as AI becomes embedded in workflows. This is not inherently negative. The core issue is inefficiency. At the same time, frontier models are likely to become more expensive, as training and inference rely on increasingly costly hardware. Properly used, tokens enable high-value thinking and creative output; poorly used, they erode AI ROI. The remedy is stronger token hygiene, which reduces cost, lowers energy consumption, and improves output quality. Increased AI usage, therefore, does not have to translate into uncontrolled cost or energy consumption growth.