What's Happening?
Cisco, through its Splunk division, released the 'Splunk State of Observability 2026: Restoring Predictability in the AI Era' report, which highlights a significant disconnect between the rapid deployment of AI agents and the ability to effectively monitor
them. The report, based on a survey of 2,050 ITOps and engineering professionals across 10 countries, indicates that the urgency to deploy AI agents is outpacing the capacity to observe and control their costs and behaviors. Over half of the respondents reported an inability to accurately control AI costs, with 60% citing unexpected billing spikes as a major operational strain. Furthermore, 83% expressed concern about the unintended consequences of AI agents, emphasizing the urgent need for 'agent observability' to monitor AI agents and establish guardrails.
Why It's Important?
The findings from Cisco's Splunk report underscore a critical challenge facing businesses adopting AI: managing the unpredictable costs and behaviors of AI agents. This issue has significant implications for corporate budgets, operational efficiency, and cybersecurity. Without effective observability, organizations risk financial unpredictability and potential security vulnerabilities as AI agents gain access to various parts of the tech stack. The report highlights the necessity for increased collaboration between ITOps, engineering, and security teams to ensure AI trust and control. For U.S. businesses, this means a potential re-evaluation of AI deployment strategies, prioritizing robust monitoring and governance frameworks to mitigate risks and maximize the value of AI investments.
What's Next?
The report suggests that observability is key to restoring predictability and confidence in the agentic AI era. Organizations are urged to implement 'agent observability' to monitor AI agent behavior, measure quality, and tie performance to business outcomes. This will likely lead to increased investment in unified observability platforms that can provide a comprehensive view of AI operations and costs. The need for closer collaboration between observability and security teams is also a critical next step, as shared data across these functions is essential for AI trust. Companies will need to develop more sophisticated 'AI tokenomics frameworks' to transform AI from an unpredictable expense into a measurable and strategic long-term business investment.
Beyond the Headlines
The 'hidden costs' of AI agents extend beyond mere financial outlays, encompassing potential ethical dilemmas, data privacy concerns, and the risk of AI 'hallucinations' or unintended actions. As AI agents become more autonomous and integrated into critical business processes, the ability to understand, explain, and control their decisions becomes paramount. This report hints at a growing need for regulatory frameworks and industry standards around AI observability and accountability. The convergence of observability and security data, as highlighted by Cisco, also points to a future where these traditionally separate domains become increasingly intertwined, driven by the complex and interconnected nature of AI systems. This could lead to new roles and skill sets required within IT departments.
