# Enterprise AI Agents: Security and Governance Shortcomings Companies are using AI agents without proper controls. - **Source:** [VentureBeat](https://venturebeat.com) - **Published:** 2026-07-24 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/487f92fa-3032-479d-b223-fb8a981f731b ## Introduction Companies started using AI agents without the necessary controls. This was revealed in five parallel surveys conducted by VentureBeat Research in June 2026. Now, these companies are retrofitting to catch up with their own standards and budgeting for it. ## Control Layers VentureBeat Research measured the five control layers needed for an agent to be trusted: identity, evaluation, cost telemetry, context layer, and orchestration. These controls determine what agents can do and how they operate. ## Security Issues 69% of companies allow their agents to share credentials. This increases security incidents. Scoped identity for each agent can solve this problem. ## Cost and Efficiency More than 80% of companies use their GPUs at 50% or less capacity. Cost and efficiency tracking are crucial parts of AI infrastructure. ## Context Layer 57% of companies trace incorrect agent answers to missing or inconsistent business context. Correct metrics and definitions are priorities for scaling agents. ## Conclusion The reliability and governance of AI agents have become critical issues for companies. Increasing controls and proper infrastructure investments can accelerate this process. ## Sources covering this story - https://venturebeat.com/technology/venturebeat-research-where-enterprise-ai-agent-governance-hasnt-caught-up — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._