Enterprises are aggressively pivoting toward agentic AI to drive autonomous IT operations, even as significant gaps in infrastructure readiness and human trust persist. According to the Riverbed Global Survey 2026 on "The State of Autonomous IT Operations," 90% of surveyed organizations intend to utilize agentic AI to achieve autonomous outcomes. While 91% of respondents report that AI investments have met or exceeded their ROI expectations, a massive disconnect remains between strategic ambition and operational reality. Most organizations are currently operating in the early stages of this transition, with only 19% of IT operations automated today. This tension between high investment confidence and low automation levels suggests that the primary hurdle for the enterprise has shifted from proving AI's value to executing it at scale.
The Gap Between AI Ambition and Operational Readiness
The transition toward autonomous IT is characterized by a stark divergence between executive goals and technical capabilities. While 76% of organizations view agentic AI as critical or very important to their future IT strategy, only 8% have achieved enterprise-wide AI-enabled operations. This discrepancy is further highlighted by the fact that only 41% of organizations believe their current environments are fully prepared for AI integration. The survey, which polled 1,200 business decision-makers and technical specialists across seven countries, indicates that the path to autonomy is currently blocked by foundational data and visibility issues.
Specifically, a "reality gap" of 79 points exists regarding visibility; while 96% of organizations recognize that multi-domain visibility is essential for AI-driven operations, only 17% have successfully unified visibility across networks, applications, endpoints, and cloud environments. Data quality also remains a significant bottleneck. Only 23% of respondents rate their data granularity as excellent, and a mere 21% describe their data quality as excellent. Without these foundational elements, the survey suggests that the move toward highly autonomous systems—which 90% of organizations expect to reach within the next two years—will remain stalled in the experimental phase.
Governance and Trust as the New IT Domains
As AI moves from experimentation into the core IT operating model, the requirement for human oversight is becoming a non-negotiable constraint. Despite the push for autonomy, 77% of organizations are hesitant to permit AI to make operational decisions without human approval. This lack of confidence is driven by several key obstacles: 55% of respondents cite security and compliance concerns, 45% point to the risk of operational disruption, and 39% identify a lack of trust in AI decisions as a primary barrier.
This hesitation is creating a new requirement for specialized management capabilities. The research indicates that 92% of respondents believe AI observability and governance will emerge as a critical new IT domain. Currently, confidence in existing frameworks is low; only 42% of respondents—and just 36% of technical specialists—report being highly confident in their organization's AI governance. This suggests that for agentic AI to move beyond human-supervised models, enterprises must first solve the problem of "black box" decision-making. The survey highlights that only 24% of organizations extensively trust AI recommendations when provided with limited human review, underscoring the urgent need for AI that is both understandable and governable.
Key Takeaways
- 90% of organizations plan to pursue human-supervised or highly autonomous IT operations within the next two years.
- A 79-point gap exists between the 96% of leaders who value multi-domain visibility and the 17% who have actually achieved it.
- 77% of enterprises remain hesitant to allow AI to make operational decisions without direct human approval.
TechInsyte's Take
In our view, the Riverbed data exposes a dangerous "readiness trap" currently facing the enterprise. CIOs are being pressured to deliver the ROI promised by AI—which 91% say is already materializing—yet they are attempting to build autonomous structures on top of fragmented, low-quality data foundations. The 79-point visibility gap is not merely a technical hurdle; it is a strategic risk that could lead to the very operational disruptions that 45% of leaders fear. We believe the next 24 months will see a shift in budget allocation away from "more AI" and toward "better data and governance." Until enterprises can bridge the gap between unified visibility and the current 17% reality, agentic AI will remain a supervised tool rather than a true autonomous agent. The winners will not be those with the most advanced models, but those with the most transparent and governed data pipelines.
Questions & Answers
How prepared are enterprises to implement agentic AI at scale?
Current readiness is low. While 90% of organizations aim for autonomous operations within two years, only 19% of IT operations are currently automated, and only 8% have achieved enterprise-wide AI-enabled operations.
What are the primary barriers preventing autonomous decision-making?
The main obstacles are security and compliance concerns (55%), the potential for operational disruption (45%), and a fundamental lack of trust in AI-driven decisions (39%).
Why is multi-domain visibility considered a critical requirement?
Multi-domain visibility is essential for AI to function across networks, applications, and endpoints. However, a significant gap exists: 96% of organizations see it as critical, but only 17% have achieved full unification.
How is the role of the IT service desk expected to change?
The service desk is expected to undergo a major transformation. By 2028, 27% of respondents predict autonomous operations will significantly reduce service desk activity, while 35% expect roles to shift from reactive support to proactive prevention.
Source: Riverbed
