Demand for AI observability is accelerating across corporate IT departments as enterprise adoption expands. A new survey from Dynatrace of 1,202 technology leaders across 12 countries in August 2026 shows that 87 percent of organizations deploy artificial intelligence agents for observability and IT operations. However, significant operational gaps remain between organizational goals and execution.
According to the findings, 44 percent of respondents target fully autonomous operations within five years, compared to 12 percent today. In contrast, human specialists still verify an average of 66 percent of automated decisions. Furthermore, only 27 percent of organizations currently track how software agents make decisions.
Adoption of AI observability and Operational Gaps
The study indicates that across nine AI observability capabilities, practical execution trails perceived importance by an average of 14 percentage points. Specifically, 43 percent of leaders consider agent decision visibility important, yet few teams actively address it. Consequently, leaders prepare systems to take actions before establishing how decisions occur.
Meanwhile, enterprise spending on infrastructure continues to climb alongside algorithmic expansion. About 78 percent of organizations expect their monitoring budgets to increase, with 96 percent identifying artificial intelligence adoption as a key driver. Additionally, 81 percent report higher compute costs from advanced models.
Rising Infrastructure Complexity and Budget Shifts
Technology complexity is growing alongside software deployments across modern cloud architectures. Exactly 43 percent of technology executives cite model and agent expansion as a leading source of system complexity. As a result, 77 percent of organizations report an urgent requirement for precise cost-allocation tracking.
In addition, system integration serves as a primary barrier to complete automation. About 31 percent of surveyed leaders identify integration complexity as their top constraint, followed by skills gaps and data validation. Notably, enterprises deploy nearly four distinct validation techniques to confirm operational outputs.
Runtime Security and Technical Exposure
Operational risks also extend directly into enterprise cybersecurity defenses. Among security respondents, 55 percent experienced a zero-day incident during the previous 12 months. However, only 45 percent maintain continuous runtime monitoring for active vulnerabilities in production.
Static vulnerability evaluations show potential flaws, whereas runtime data verifies active exposure. Consequently, security teams are deploying automated monitoring specifically for agents and models. This operational shift provides practical evidence to prioritize high-risk system updates.
Open Standards and Future Autonomy
Open frameworks are gaining significant traction across enterprise IT environments. The study shows 77 percent of organizations moved OpenTelemetry past evaluation stages, while 20 percent established it as an operational standard. Furthermore, 37 percent cite agentic tracking as their primary motivation for adoption in apps & software management.
Deploying AI observability tools enables organizations to verify automated actions across hybrid systems. As enterprises aim for autonomous operations, linking technical telemetry with real business outcomes provides necessary governance for ongoing growth.
