Progress Software adds agentic RAG with Teams and MCP support
Progress Software (Nasdaq: PRGS) has released a set of new capabilities for its Agentic RAG platform, extending the product's retrieval-augmented generation tooling with a native Microsoft Teams application, a Smart Agent for autonomous multistep retrieval, and a WordPress plugin for improved content ingestion. The Burlington, Massachusetts-based company says existing customers inherit all three features without reindexing or architectural changes to their current deployments.
The update is centred on what Progress calls an Agentic Knowledge Layer, a shared retrieval foundation that lets AI agents pull from both pre-indexed enterprise content and live business systems in a single query. Native support for Anthropic's Model Context Protocol (MCP) in the Smart Agent allows agents to reach live data sources at query time without requiring custom integrations for each connected system, which Progress positions as a way to reduce bespoke orchestration work for engineering teams.
What the release adds
The Teams application, available through the Microsoft Store, brings source-cited AI responses directly into Teams chats and channels, preserving existing permissions and governance controls. The Smart Agent breaks complex queries into sub-questions, retrieves information across indexed content, MCP-connected applications and web sources, and evaluates results iteratively until an adequate answer is assembled. The WordPress plugin gives administrators finer control over how CMS content is indexed, preserving metadata and taxonomy to improve retrieval fidelity.
John Ainsworth, EVP and General Manager of Application and Data Platform at Progress, said the next phase of enterprise AI "will be defined by AI that can navigate complex questions, work across organisational knowledge and deliver answers where people already work." The company did not disclose the number of active Agentic RAG deployments, customer names, or pricing changes associated with the new release.
Market context
Agentic RAG sits at a crowded intersection of the enterprise AI stack. Microsoft Copilot, Glean, Elastic and a growing number of well-funded startups are all pursuing variations of the same proposition: connecting foundation models to proprietary enterprise knowledge without requiring ground-up rebuilds of existing data pipelines. Adoption of MCP as a standard for agent-to-system connectivity is gaining traction across the sector, with several platform vendors adopting it over custom tool-calling schemas in recent months.
For Progress, the release continues a strategic pivot toward AI infrastructure that the company has signalled through product positioning and investor communications over the past year. Its installed base spans a wide range of enterprise environments, giving it a distribution advantage if it can convince customers to consolidate RAG and agent orchestration onto a single governed platform rather than assembling point solutions.
Regulatory and governance read-across
Enterprise adoption of agentic AI is increasingly shaped by governance concerns rather than technical capability alone. The EU AI Act's obligations for general-purpose AI systems are phasing in through 2027, and several financial services and public-sector buyers are requiring auditability, source citation and access-control enforcement as baseline requirements before any agent deployment. Progress's emphasis on governance, permission inheritance and source-cited responses appears calibrated to address exactly these procurement criteria, though the company has not published third-party audit results or compliance certifications specific to the Agentic RAG platform.
The immediate milestones to watch are customer case studies that demonstrate measurable retrieval accuracy or adoption gains, and whether MCP support broadens to cover the full range of enterprise systems that buyers typically need to connect.