Red Hat has taken a significant step toward integrating artificial intelligence into enterprise IT automation by opening its Ansible Automation Platform to AI agents. On Tuesday, the company announced the general availability of its Model Context Protocol (MCP) server for Ansible, enabling external AI agents to connect to the platform. Additionally, Red Hat previewed a new automation orchestrator designed to keep AI actions under tight control by routing them through deterministic, human-approved playbooks.
The move addresses a growing demand among enterprises to leverage AI for workflow automation while maintaining strict governance. Recent high-profile incidents of AI agents performing unauthorized actions have underscored the need for safeguards. Red Hat's approach aims to balance the flexibility of AI with the predictability of established automation practices.
How the MCP server and orchestrator work
The MCP server provides a standardized interface for AI agents to interact with Ansible Automation Platform. This allows tools from various providers—including models from Google, Anthropic, OpenAI, and any OpenAI API-compatible models—to request automations. The orchestrator, currently in technology preview, ensures that any action proposed by an AI goes through a pre-vetted playbook. If the AI suggests an action that does not match an existing playbook, a human must review and approve it before execution.
Red Hat has also enhanced Ansible to support retrieval-augmented generation (RAG), allowing enterprises to inject their own context—such as internal policies, maintenance windows, and infrastructure rules—into the AI models. This enables the AI to make more informed and compliant suggestions without deviating from organizational norms.
Why guardrails matter
“AI is unpredictable,” said Sathish Balakrishnan, vice president and general manager of the Ansible business unit at Red Hat. “When you suddenly put AI into your production environment and ask it to change it, you've seen the articles about how a company lost its database.” By relying on playbooks that are testable, repeatable, and deterministic, Red Hat minimizes the risk of AI-driven errors. The playbooks also reduce reliance on expensive token calls to large language models during actual automation runs, as noted by Balakrishnan: “Why would you use AI just to patch a machine? We all know tokens are expensive. We know the best way to patch a machine—why call an AI to do that when you already have a playbook that's been in use for ten years?”
Industry analysts have echoed the need for caution. Paul Nashawaty of Efficiently Connected warned that connecting agents to highly privileged automation systems can lead to catastrophic failures if not properly controlled. “The security concerns are very real,” he said. “If those agents are connected to highly privileged automation systems, the blast radius can become enormous, including accidental production outages or destructive actions.”
Use cases and best practices
Red Hat and other experts see the strongest AI applications in areas such as AI-assisted troubleshooting, compliance remediation, developer self-service, and human-approved workflow execution. For example, developers can request environments in natural language, or AI systems can correlate alerts and suggest fixes for operations teams. In all cases, human oversight remains a key component.
IDC analyst Jevin Jensen noted that natural-language interfaces have been long-awaited by platform vendors. “This really broadens the use and value of the platform to new users and improves efficiency of existing users,” he said. However, he stressed the importance of good governance, especially role-based access control, to limit the blast radius of any AI action.
Additional enhancements in the latest Ansible release include the ability for administrators to delegate trigger rights to end users—such as factory floor managers who can schedule updates during low-production windows—and support for multiple events to trigger the same playbook, reducing duplication.
Background and significance
Ansible is an open-source IT automation tool widely used for configuration management, application deployment, and task orchestration. Red Hat has been steadily expanding its AI capabilities, having previously integrated IBM's WatsonX Code Assistant. The new MCP server and orchestrator represent a more open approach, allowing any AI agent to connect while retaining strict control mechanisms. This is particularly important as enterprises experiment with generative AI in production environments.
The broader trend in IT automation is toward “human-in-the-loop” AI systems that augment rather than replace human decision-making. Red Hat's announcement aligns with that philosophy, providing a framework where AI can suggest actions but not execute them without validation. This approach is likely to become a standard for enterprise automation as companies seek to balance innovation with risk management.
In summary, Red Hat's latest updates to Ansible Automation Platform open the door to AI agents while ensuring that their actions remain predictable and safe. By combining the power of large language models with deterministic playbooks and human oversight, the platform offers a pragmatic path for enterprises to adopt AI-driven automation without exposing themselves to unnecessary risk.
Source: Network World News