In a recent analysis, John Encizo of Lenovo highlights the risks associated with deploying agentic artificial intelligence (AI) in organizations. According to Lenovo's CIO Playbook 2026, only 10% of organizations are prepared for large-scale deployment of agentic AI, while Deloitte’s State of AI report indicates that just 25% have successfully transitioned at least 40% of their pilots into production systems. This suggests a significant gap between ambition and readiness.
Encizo emphasizes the importance of understanding how workflows break before scaling up AI deployments. Organizations that start small, move successful pilots to full-scale production, and thoroughly map out their processes are better positioned for success. He also notes that many companies lack the necessary infrastructure or best practices to manage agentic AI effectively.
A key risk in multi-agent systems is the compounding effect of small errors across interconnected agents, which can escalate into major issues if not addressed promptly. Most organizations have yet to implement robust circuit breakers to mitigate such risks.
Encizo observes that while some U.S.-based companies are moving rapidly toward implementing numerous AI projects, Canadian firms tend to adopt a more measured approach. This balanced strategy may be better suited for the current phase of technological
Read the full article at Digital Journal
Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.
![[AINews] The Unreasonable Effectiveness of Closing the Loop](/_next/image?url=https%3A%2F%2Fmedia.nemati.ai%2Fmedia%2Fblog%2Fimages%2Farticles%2F600e22851bc7453b.webp&w=3840&q=75)
