Enterprise AI agent deployments hampered or abandoned due to fragmented data infrastructure

According to research from Cloudera and Google/MIT reported by CIO, enterprises show fervent interest in agentic AI but deployments continue to be hampered or abandoned largely due to data access, context, and governance issues. Google/MIT noted that organizations cannot ignore their data systems as they attempt to scale agentic AI, indicating that AI agents require fundamentally different data infrastructure than systems built for human operators.

Update (2026-08-18): According to CIO, Ben Schein, chief AI and analytics officer at Domo, reports that organizations find AI agents easy to pilot but difficult to scale; costs concentrate in t...