Cisco report examines global AI readiness and gaps in infrastructure and security

Pacesetters achieve more widespread results than their peers because of this approach: 90% report gains in profitability, productivity, and innovation, compared with 60% overall.

Jeetu Patel, Cisco’s President and Chief Product Officer. 

Cisco has released its third annual AI Readiness Index, a global study analyzing how organizations are adopting and managing artificial intelligence technologies. The survey covered more than 8,000 business and technology leaders across 30 countries and 26 industries.

The report shows that a small group of organizations, about 13% of respondents,  consistently demonstrate higher performance in moving AI projects from pilot to production and achieving measurable business outcomes. These companies, referred to in the study as “pacesetters,” are more likely to have formal AI roadmaps, investment strategies, and scalable infrastructure designed for AI workloads. According to Cisco, 98% of these organizations have networks prepared to handle AI growth and complexity, compared to 46% among other companies.

The research also highlights two key developments shaping enterprise adoption: the growing use of AI agents, systems that can independently perform tasks, and the increasing challenge of AI infrastructure debt, which refers to legacy systems and deferred upgrades that limit scalability and performance.

Most organizations plan to deploy AI agents soon: 83% expect to do so, and nearly 40% anticipate these tools will work alongside employees within a year. However, many companies report insufficient infrastructure, with more than half indicating their networks cannot scale to support complex AI systems. Only 15% describe their networks as adaptable to rapid changes in data volume or computational demand.

Cisco’s study also identifies infrastructure and data management challenges as potential risks to long-term value creation. About 62% of companies expect workloads to increase by more than 30% in the next three years, while 64% face difficulties in centralizing data. Additionally, only 26% report adequate GPU capacity to handle advanced AI models, and fewer than one in three organizations have systems in place to detect or prevent AI-specific security threats.

The findings suggest a widening gap between AI ambitions and operational readiness. While some organizations are advancing quickly, others may face constraints due to limited infrastructure, governance frameworks, or security measures needed to support large-scale AI adoption.