Microsoft Azure leaders say agent confidence hinges on business context

Enterprise investment in AI is booming, and Gartner is calling 2026 an “inflection year” for organizations to align AI projects with strategic business objectives. The subtext is obvious: executives are getting pressed to prove ROI, and teams need a concrete mechanism to turn AI spend into measurable outcomes. That is why agentic AI is showing up everywhere from planning to operations. But a new MIT Technology Review report, produced by Insights (its custom content arm), argues the real bottleneck is not whether agents can do tasks. It is whether they have the business context required to do them safely, reliably, and securely.

The clearest statement in the report comes from Jeremy Winter, corporate vice president and chief product officer at Microsoft Azure Platform. His core point is that “as we design agents to operate within the same operational boundaries, identity systems, and governance models that teams already use, they start to behave more like the systems organizations already trust.” In other words, confidence does not come from bigger promises. It comes from mapping agent behavior onto the guardrails enterprises already understand.

To ground that claim, the report draws on research and ranks 101 tasks across AI, data, and cloud workflows based on respondents’ confidence in agents acting on their behalf. It also examines how tech teams view both the opportunities and challenges of agentic AI, including what the potential means for their careers. The headline finding is straightforward: among technology experts, teams are exceedingly confident about using agentic AI across a significant amount of AI, data, and cloud tasks. Where readiness drops, the report says, is largely due to a lack of business context being supplied to agentic systems.

Why does context matter so much? The report’s logic is practical. The more complex the task, the more reasoning capability an agent requires, and the greater its need for business context. That makes context generation a foundational requirement, and the report places it at an early stage of development, especially in situations where enterprise data is difficult to wrangle and connect into the agent lifecycle at the speed and quality developers and executives need. This is not a niche engineering problem. It directly connects to risk management, auditability, and operational continuity, because enterprises cannot delegate decision-making without confidence the agent will operate within safe, reliable, and secure boundaries.

Human oversight shows up as a key factor of success. Even when confidence is high, teams are not treating agent delegation as a blind handoff. Knowing that tech teams are in a pivotal position to lead this transformation, the experts the report interviewed expect agent confidence to accelerate as experience with agents deepens and business environments mature. The trust-building mechanism is also explicitly tied to execution: if agents are designed to work within the same identity systems, governance models, and operational boundaries already used by teams, then the behavior becomes more familiar, and confidence grows.

The report also paints a domain-by-domain picture of where confidence is highest. It says confidence is surging for measurable tasks and growing in areas of complex judgment. Technology experts overwhelmingly believe agents help with everyday work, including streamlining processes, improving performance, and reducing repetitive tasks. The highest confidence is for processes like generating reports and boilerplate code. But the report also highlights where opportunity concentrates: tasks involving multistep workflows and advanced reasoning to make decisions. This matters for leadership because “everyday work” is easy to pilot, while multistep, reasoning-heavy workflows are where ROI either compounds or collapses.

Data workflows are described as the breakthrough domain for agentic systems. Teams trust agents most where structure can provide a reliable foundation for decisions, including data quality monitoring, visualization anomaly detection, real-time data stream monitoring, and data profiling. The report frames this as a zone where domain experts closest to the point of data generation can provide context so agents can act and deliver trusted outcomes. For boards and C-suites, that is a strategic hint: the fastest path to credible agent deployment likely runs through the data systems where context can be standardized, monitored, and governed, rather than through the messiest parts of enterprise knowledge where context is hardest to assemble.

Zoom out and the incentives become clearer. In the last 18 months, tech teams, including engineers, developers, architects, and other practitioners building, deploying, and continually improving infrastructure and applications, are clearly putting agents to work. Meanwhile, the report also notes that IT infrastructure costs are projected to grow two to three times by 2030 even as budgets remain unchanged, citing McKinsey. That squeeze makes automation tempting, but it also makes trust non-negotiable. If agent confidence depends on business context and oversight, then the real competitive advantage is organizational: how quickly enterprises can connect data to the agent lifecycle, and how effectively they can embed agents within existing governance structures.

For executives looking at similar initiatives, the stake is simple. If you can only prove ROI on narrow, structured tasks, you may get pilots. If you can scale agentic workflows with reliable context, you can get repeatable outcomes. And in an “inflection year” framed by strategic alignment and ROI pressure, the difference between pilot and transformation is whether your agents can earn trust, not just trigger automation.