A recent study by Harvard Business Review’s Analytic Services highlights a significant challenge facing businesses as they strive to implement agentic artificial intelligence (AI) solutions. While nearly all organizations recognize the critical importance of AI, only a minority have the necessary infrastructure to support it.
The survey reveals a stark contrast between enthusiasm and readiness. An overwhelming 96% of senior executives believe agentic AI will be pivotal in the coming years. However, only 23% feel equipped with the infrastructure to bolster these capabilities. Furthermore, nearly half of the companies deploying agentic AI indicate that the associated infrastructure costs have already surpassed their initial expectations. Some firms anticipated eliminating security triage teams, only to learn they need a substantial infrastructure investment.
A significant part of the challenge lies in the telemetry foundation. As AI becomes more integrated, data systems need to evolve. Traditional dashboards are insufficient for managing the data volume and speed required by agentic AI, leading to potential operational black boxes. Ryan Kurt, CEO at The AI Lab, stresses the necessity of robust infrastructure: “Without the right infrastructure, you’ll hit a ceiling.”
IT and security teams are particularly feeling the strain as they attempt to manage next-generation AI on outdated systems. Unlike general AI, agentic AI requires real-time analysis across multiple systems. Legacy tools may capture previous actions, but they fail to provide the immediate “why” necessary for modern AI-driven decision-making.
Organizations at the forefront of this AI transition stress the importance of updating their data infrastructure. Instead of solely purchasing more AI tools, they are revamping their data ecosystems to seamlessly integrate machine and human data for more effective AI reasoning. Adopting flexible, interoperable platforms is becoming essential.
Harvard Business Review’s survey also highlights several key points. As agentic AI is deployed, 76% of organizations report increased telemetry data volumes, with some experiencing more than twice the volume. There are financial implications, with 82% expecting increased infrastructure costs. Challenges also include privacy concerns, talent shortages, and data architecture inadequacies, which slow AI project progress.
The research surveyed 351 business leaders from North America, Europe, and Asia Pacific in early 2026. Clint Sharp, CEO at Cribl, noted, “Data is growing at a 30% CAGR, budgets are not, and now AI agents are multiplying that problem.” He underscores the urgency for companies to modernize their infrastructure to remain competitive as AI continues to shape the future.
