OpenAI updated Agents SDK: enterprise security and multi-agent scenarios

OpenAI has updated its Agents SDK — a set of tools for developing autonomous AI agents — with an emphasis on security and enterprise-scale capabilities. The company has expanded the framework's functionality in response to growing demand from large enterprises that want to build complex agent systems with predictable and controlled behavior. The Agents SDK, first introduced in early 2025, was initially oriented toward developers prototyping agents.

The new update shifts focus toward production-ready solutions: agent scenario execution tracing has been improved, and guardrails capabilities have been expanded — validation mechanisms for incoming and outgoing data. Enterprise customers can now more precisely control what context an agent receives and what it returns to users, preventing unwanted or unsafe responses. One of the key innovations is improved support for multi-agent orchestrations.

The SDK received more flexible handoff mechanisms — transferring control between agents within a single workflow. This is particularly important for complex enterprise scenarios, where a task is sequentially processed by several specialized agents: one searches for information, another analyzes it, a third formulates the final response. The new version makes such chains more reliable and transparent for developers.

OpenAI also reported improvements in observability tools. Integration with tracers allows teams to see in real time what steps an agent took, which tools it called, and at what stage errors occurred. This is critically important for auditing in regulated industries — banking, healthcare, legal sector — where every automated solution must be explainable and documented.

Interest from big business in agent AI systems continues to grow. According to analysts, the number of enterprise pilots with AI agents more than doubled over the past year. Companies are looking for ways to automate multi-step workflows — from processing customer requests to managing IT tickets and analyzing documents.

OpenAI positions the updated SDK as a tool that lowers the entry barrier for enterprise developers who until now have had to build agent infrastructure from scratch. Competition in the agent framework market is intensifying. Anthropic is developing the Claude Agent SDK with an emphasis on constitutional safety.

Google offers an Agent Development Kit with native integration into Vertex AI. Microsoft is pushing AutoGen and Semantic Kernel, while independent frameworks like LangChain maintain their share among open-source developers. OpenAI's update is part of the race for enterprise sector trust, where security and standards compliance requirements are significantly stricter than in consumer applications.

For the market as a whole, this is a signal: agent systems are ceasing to be experimental technology and becoming a production standard. Major vendors are investing not just in launching agents, but in tools for their long-term management — with versioning, monitoring, and compliance with corporate security policies.