Enterprises winning with AI agents are limiting how much the agents can do alone
For much of the past two years, the general belief in enterprise AI has been that more autonomy equals better performance. Build agents that can plan, decide, and act across multi-step workflows, and give them as much room to run as possible. That assumption is now being tested at scale, in real production environments — and in a lot of deployments it's failing. The companies that end up…
Enterprise AI agents are seeing success when they are given specific responsibilities and operate within clear rules, rather than being given maximum autonomy. Two key metrics illustrate the current state of agentic AI as of mid-2026. According to Gartner, over 40% of agentic AI projects running today are not expected to survive until 2028 due to escalating costs, unclear business value, and inadequate risk controls.
Similarly, a 2026 AI Trust Maturity Survey by McKinsey shows that only about 30% of organizations have reached a maturity level of three or higher in governance and agentic AI controls specifically. This gap between capability and control is shifting the competitive landscape. The initial race focused on deploying the most autonomous agents fastest, but now the emphasis is on trust.
Enterprises are prioritizing the approval of agents for production, ensuring they are manageable and compliant with risk, legal, and compliance requirements. Full autonomy often breaks down in real-world production environments due to integration complexity, scalability issues, and lack of transparency. Autonomy and accountability have opposing directions; an agent capable of independent decision-making also has harder-to-trace decisions.
This lack of transparency can lead to regulatory breaches in sensitive areas such as finance, compliance, manufacturing, or healthcare. Integration complexity is a leading cause of project cancellations, as it requires reconstructing existing workflow decision points, approval chains, and audit trails to accommodate autonomous agents.
However, leading enterprises are not giving up on AI. They are restructuring how autonomy is distributed within their systems. Successful organizations follow four key patterns: they deploy narrow-scope agents with specific responsibilities, establish human checkpoints before high-stakes actions, ensure decision traceability as a design requirement, and focus on data sovereignty for active governance.
These strategies allow enterprises to balance capability with control, ensuring their AI agents perform effectively and safely within real-world production environments.
Written by urgent.news from VentureBeat's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.