Agentic AI ROI: New Framework for Value & Risk

Traditional methods for calculating the return on IT investments are not working for agentic AI. These AI systems execute multi-step tasks autonomously and make real-time decisions, meaning their value is nonlinear. Forty-two percent of organizations worldwide already find it difficult or impossible to assess the ROI of their digital and AI investments. Agentic AI amplifies these challenges. Its value is shaped by adoption quality, autonomy boundaries, feedback loops, and data quality. Costs are also complex, driven by factors like orchestration and governance, not just licensing and API calls. Failures in agentic systems can propagate, which static risk models don't account for. The International Data Corporation has developed the Agentic AI Business Value Maximization Framework, a six-pillar approach with governance as a core element. This framework helps organizations define where autonomy is permitted and constrained, moving beyond vague productivity goals. This matters because understanding the true value and risks of agentic AI requires a completely new measurement approach.