Agentic AI Deployments Stalled by Data and Governance

A significant number of enterprises are now deploying autonomous AI agents in live business operations, but widespread challenges with data infrastructure, governance, and organizational readiness are stalling progress and forcing a strategic rethink. According to a new industry survey, 32% of organizations are currently using agentic AI in production, a modest increase from 29% the prior year. However, among those with live deployments, 77% report stalled projects and 61% cite project abandonment as a problem, indicating that reaching production is often where the most difficult challenges begin.

Agentic AI refers to systems that can pursue goals, make decisions, and take actions-such as routing IT tickets or executing workflow steps-with limited human prompting. This shift from AI that assists to AI that acts introduces new operational complexities. The primary barriers to success, as identified by IT leaders, are a skills gap and lack of organizational readiness (69%), concerns about large language model reliability (68%), and data infrastructure and quality issues (66%). Governance, risk, and compliance challenges are also a major concern for 65% of respondents.