91 Percent Use AI. Only 33 Percent Trust AI Agents With Real Work: Report

A new report suggests enterprise enthusiasm for AI has not yet translated into confidence that autonomous agents can do real business work.

Most organizations have embraced artificial intelligence, but relatively few are willing to let AI perform meaningful work on their behalf, according to a new report from Tempo Software.

The company's 2026 State of AI in Portfolio Management Report found that 91 percent of organizations are piloting or actively using AI in project delivery. However, only 33 percent currently delegate work such as coding, documentation, or quality assurance to AI agents. Tempo's report also found that 42 percent of leaders cannot associate AI spending directly with return on investment.

The report suggests a hesitancy among business leaders to leave critical tasks up to AI agents, despite an ongoing race to adopt AI in business operations.

"The results show that using AI is not, by itself, an advantage," Tempo Software CEO Vic Chynoweth said in a news statement. "What matters is whether organizations can put agents to work, and govern that work alongside their people, investments and strategic priorities."

For midmarket IT leaders, the findings align with a trend MES Computing has been tracking throughout 2026: organizations are widely embracing AI tools but struggle moving AI projects into production.

AI projects often stall because of integration, infrastructure and data issues, even after successful proof-of-concept deployments.

Agentic AI adds more complexity. In recent MES Computing coverage of emerging AI-powered procurement tools, the key question was not whether organizations could properly govern those decisions. As that article noted, if an AI agent recommends one supplier over another, CIOs may need visibility into how these recommendations were made and how those decisions could be audited later.

Tempo's findings echo similar concerns.

The report revealed that 39 percent of leaders cannot distinguish AI-generated work from human-produced work within their current tools. At the same time, 42 percent said they cannot directly connect AI spending to measurable business outcomes.

Those numbers may shed some light on why organizations remain reluctant to hand critical work to AI agents despite overall enthusiasm for the technology.

Midmarket organizations face related governance challenges: understanding how AI is being used, how underlying models change over time, and what impact those changes may have on business outcomes. As more agents start taking actions inside a business's operations, visibility, accountability, and auditing become crucial.

That hesitation Temp found with its report comes despite evidence that organizations deploying AI agents may be seeing operational benefits. Tempo found that organizations with AI agents in production report faster delivery, with 71 percent of projects completed within six months compared with 55 percent at organizations without deployed agents.

Tempo's report findings are based off a a survey conducted in June 2026 among 300 senior project, portfolio and PMO leaders at organizations with between 200 and 4,999 employees and annual revenue ranging from $100 million to $5 billion.