OutSystems’ Agent Workbench reaches general availability, helping enterprises streamline operations through agentic AI (www.techradar.com)

OutSystems announced the general availability of Agent Workbench at its ONE Conference in Lisbon — a low-code platform for building, orchestrating and governing agentic AI across enterprise workflows and legacy systems. The release packages early-adopter validation (efficiency gains in customer escalation, automated routing and loan underwriting), an agent marketplace for sharing prebuilt workflows, and features aimed at addressing enterprise pain points around security, observability and integration. OutSystems positions the product as a way to accelerate operational transformation while maintaining control over agent behavior and auditability. Key technical additions include support for the Model Context Protocol to link agents to enterprise systems and external tools, plus expanded compatibility with a broad set of LLM providers (OpenAI, Anthropic, Google, IBM and others). Agent Workbench is available in ODC and OutSystems Personal Edition (testing); pricing starts at €20K with a promotional waiver if you deploy an agent before end of 2025. For AI/ML teams, the product promises faster agent development via low-code composition, multi-model flexibility for model selection or fallbacks, and built-in mechanisms for governance and integration — practical building blocks for scaling agentic workflows in regulated or legacy-heavy environments.