Agentic AI hits critical mass, but only 18% of organizations track its ROI

Agentic AI has cleared the threshold from pilot project to professional expectation — but the financial discipline required to justify that shift is lagging well behind, according to new research from Thomson Reuters.

Adoption outpaces accountability

Thomson Reuters reports that 77% of professionals expect agentic AI to sit at the center of their daily workflows by 2030, a figure that underscores just how quickly the technology has moved from speculative to strategic.

Yet the same data reveals a striking accountability gap: only 18% of organizations currently track the return on investment of their agentic AI deployments. That means more than four in five firms are committing resources to a technology without a framework for measuring whether it is paying off.

The disconnect matters because agentic systems — which autonomously plan, sequence, and execute complex tasks — carry operational risks and infrastructure costs that differ meaningfully from earlier generations of AI tooling. Without ROI tracking, organizations lack the data needed to scale responsibly or to course-correct when deployments underperform.

Productization accelerates in consulting

While measurement practices remain immature across the broader market, some professional services firms are moving decisively to commercialize agentic capabilities. Consulting group Sia expanded its Agent Store from 50 to over 400 agents available for direct client consultation, according to Thomson Reuters, representing an eightfold increase in its catalog of deployable AI agents.

The move signals a structural shift in how consulting firms position AI: not solely as an internal efficiency tool, but as a billable product delivered directly to clients. That model compresses the distance between AI development and client value, but it also raises questions about how quality, reliability, and outcomes are governed across a rapidly expanding agent portfolio.

For enterprise buyers, the growth of agent marketplaces creates both opportunity and complexity. Procurement teams accustomed to evaluating software licenses or consulting engagements now face a hybrid category that blends elements of both — and that operates with a degree of autonomy that traditional vendor assessments were not designed to scrutinize.

The measurement gap as a strategic risk

The 18% ROI-tracking figure is likely to draw attention from CFOs and boards as agentic AI budgets grow. Organizations that cannot demonstrate financial returns face mounting pressure to justify spend, particularly in an environment where technology investment decisions are subject to increasing scrutiny.

Thomson Reuters' findings suggest the industry is at an inflection point: adoption has achieved critical mass, but the operational and financial infrastructure to support that scale has not. Firms that close the measurement gap first will be better positioned to make the case for continued investment — and to identify where agentic deployments are generating genuine value versus where they are simply generating activity.

For technology vendors, system integrators, and enterprise buyers alike, the next competitive frontier may be less about building more capable agents and more about building credible frameworks for proving what those agents are actually worth.