Agent confidence on the technical frontier

Enterprise investment in AI is experiencing a boom, with 2026 anticipated as a pivotal year for organizations to align AI projects with strategic business objectives. Executives and technology leaders are increasingly turning to agentic AI to drive measurable financial outcomes and substantial ROI. This focus is particularly keen within the tech function, where IT infrastructure costs are projected to significantly rise by 2030, even as budgets remain stagnant. Teams of engineers, developers, and architects are actively deploying agents to build, improve, and manage infrastructure and applications.

The ultimate goal of agentic AI extends beyond mere task automation; it aims to manage and coordinate entire workflows, fostering collaboration between humans and agents to achieve business objectives. This necessitates a high degree of confidence in agents' capabilities to perform tasks safely, reliably, and securely, especially given the inherent risks in automated decision-making. Research indicates that technology experts are highly confident in using agentic AI for a significant portion of AI, data, and cloud-related tasks.

However, the readiness of agents for more complex tasks is often hampered by a lack of sufficient business context. The more intricate a task, the greater the reasoning capabilities an agent requires, and the more critical it becomes to provide relevant business context. The development of robust context-generation capabilities for agents is still nascent, particularly when enterprise data is challenging to integrate into the agent lifecycle at the necessary speed and quality. Human oversight remains a key factor for successful agentic AI deployments.

Experts anticipate that confidence in agents will accelerate as experience deepens and business environments mature. Integrating agents within existing operational boundaries, identity systems, and governance models will further enhance trust. This approach allows agents to operate in a manner consistent with systems organizations already rely on.

Confidence in agents is notably high for measurable tasks and is growing for areas requiring complex judgment. Technology experts widely believe that agents can streamline processes, improve performance, and reduce repetitive tasks. This confidence is particularly strong for tasks like generating reports and boilerplate code, and there is significant potential for agents to manage multi-step workflows and advanced reasoning for decision-making. Data workflows are emerging as a breakthrough domain, where structured data can provide a reliable foundation for agent decisions, especially in areas like data quality monitoring and real-time data stream monitoring.