State of Observability 2026 Reveals 55% of IT Leaders Can't Control Agentic AI Spending

Agentic AI Creates Unprecedented Predictability Crisis

Cisco has released its State of Observability 2026 report, exposing a fundamental shift in how organizations manage AI systems1

1

2

2

. The research surveyed 2,050 ITOps and engineering professionals across 10 countries and 16 industries, revealing that 83% are concerned about the unintended consequences of AI agents. Unlike traditional systems that follow predictable routes, agentic AI operates with inherent non-determinism, making decisions autonomously without human intervention at every step. This autonomy has shattered a long-held assumption: that understanding a system once means understanding it for the future. Agents handed a goal will find their own route to it, and the route they pick today may not be the one they pick tomorrow.

The AI Tokenomics Crisis Exposes Financial Control Gap

The report reveals a striking financial predictability problem: 55% of IT leaders cannot control AI agent spending. When organizations run dozens of agents across multiple models, AI costs and usage become convoluted. Surprisingly, AI infrastructure was named as the top observability cost driver by only 6% of respondents. The real culprit is far more familiar: excessively verbose logs topped the list. Organizations have spent years collecting data just in case, sometimes to the point where it breaks budgets. Nearly half of respondents admit telemetry data isn't a key factor in their organization's strategic planning. Strategic planning without operational telemetry means businesses are essentially guessing about their future direction.

Agent Observability Emerges as Critical Business Requirement

As organizations add AI agents into workflows like software development and customer experience, an incorrect agent action can evolve into a critical business failure. Unlike monitoring traditional applications, visibility into AI agent behavior requires a new class of behavioral metrics tied to business outcomes and agent decision-making. The survey respondents feel the weight of this challenge: 83% in the global survey and 81% in India express concern about unintended consequences. The answer to this anxiety is visibility. By bringing agent behavior into observability practice, organizations can turn unpredictable actions into empirical, governable data. Observability is what makes AI spend visible, attributable, and controllable in an era where AI is exacerbating a data volume problem for many enterprises.

Leaders Separate Through Three Connected Practices

A group of top-performing organizations emerged as leaders who are achieving concrete business outcomes with their observability practices. These leaders follow three habits practiced together: they put AI agents to work fixing and preventing issues routinely rather than experimentally, they expand the paradigm of observability to govern agent behavior and enforce guardrails around model quality, cost, usage, security, and privacy, and they embed telemetry directly into business planning. This combination changes their posture fundamentally. Leaders are 2.4 times more likely than non-leaders to say the ROI of their observability platforms significantly exceeds expectations. Additionally, 54% of leaders say observability has yielded at least $5 million in annual business revenue versus 28% of non-leaders.

Security and Observability Teams Converge for AI Trust

As AI agents gain more access to operations in every layer of a business, monitoring them requires collaboration between ITOps, engineering, and security teams, breaking down walls that have historically kept them apart. "Most companies will soon have far more agents at work than people," said Jeetu Patel, President and Chief Product Officer at Cisco. "Managing agents at scale requires reimagining both security and observability. Have our agents been compromised? Are they doing their job well? In the world of agentic AI, these questions are inseparable, and they're moving at machine speed." Leaders are realizing these benefits: 94% say monitoring AI has increased collaboration with security teams compared to 80% of non-leaders, and 86% are more likely to accurately trace performance issues back to their security root causes in contrast with 71% of non-leaders.

Advanced Organizations Master Financial Predictability

Leaders have shifted attention to where problems now originate. Nearly every leader at 94% is moderately or very confident in isolating edge and branch network issues, compared to roughly two-thirds of everyone else. For AI-related issues, 68% of leaders report confidence versus 56% of peers. Leaders are far more likely to accurately attribute AI costs, with 91% of leaders versus 69% of non-leaders pinpointing which agents, models, and workloads drive spend. Kamal Hathi, SVP and GM at Splunk at Cisco, emphasized: "We believe observability is the essential bridge to restoring predictability and confidence to the agentic era. By providing the necessary guardrails and incorporating observability into an AI tokenomics framework, leaders can transform AI from an unpredictable expense into a measurable and strategic long-term business investment." Indian organizations stand out as advanced for their agent observability capabilities, with 51% calling their practice very effective at monitoring AI agents.