Your employees are building AI agents. Do you know what they’re doing?
Two-thirds of knowledge workers are building AI agents. One in seven without asking permission. IT leaders have no visibility.
Why it matters
Shadow agent creation is accelerating across enterprises as employees automate work faster than IT can govern it. The real risk isn't rogue agents—it's sanctioned ones connecting to company systems and data without controls. Governance frameworks that enable experimentation while maintaining visibility are becoming table stakes.
The key facts
9 to know66% of knowledge workers using AI also use agents (Howdy.com survey)
16% of knowledge workers have built their own agents
25% of employee-created agents have been deactivated due to inconsistency or failure
7% of knowledge workers have deactivated agents that went rogue
Employees building agents often do so without IT approval or organizational sanction
Primary use case: automating manual, repetitive tasks
Risk: agents accessing CRM data, updating records, moving information between systems without enterprise controls
AppDirect model: 'proper digital playground' with full IT visibility into usage, costs, and data access
Failing silently vs. failing and learning is the critical distinction for IT governance
Go to the source
CIOcio.com
Publisher excerpt: Eager to incorporate agentic AI into their work, employees are taking matters into their own hands, creating their own AI agents, regardless of organizational approval, and a not insignificant number of those creations could have devastating consequences if IT leaders don’t get ahead of the trend.…