Amazon Web Services proposes new framework for agentic automation value

Amazon Web Services (AWS) has unveiled a new approach to calculating the business value of agentic automation, a form of software that can reason and adapt to complete tasks. The company argues that the traditional method of justifying automation investments, based primarily on saved hours, is insufficient for capturing the full value created by these advanced agents.

The traditional return on investment (ROI) model, designed for stable, high-volume, rule-based tasks common with robotic process automation (RPA), falls short. This model often overlooks the costs associated with maintaining automation as processes change, handling exceptions, and the ongoing human oversight required. Furthermore, the freed-up capacity may not always translate into direct profit and loss gains as new backlogs can emerge.

AWS points to research by McKinsey, which suggests that successful AI transformations often follow a "1:3:5 pattern." This pattern indicates that for every dollar invested in agentic technology, three dollars should be spent on process redesign and five dollars on capability building and adoption. However, many organizations invert this ratio, potentially leading to failed implementations.

The new framework is intended to equip leaders of AI Centers of Excellence (AI CoE) with the tools to construct a more comprehensive business case. It aims to help them accurately measure the value delivered by agentic automation and identify the most strategic workflows for its implementation.

This initiative addresses the growing adoption of agentic automation on AI CoE roadmaps, providing a method to better quantify its impact beyond simple time savings.