Progress Agentic RAG Adds a Smart Agent Plus Native Teams and WordPress Support

Three New Capabilities Land Today and Every Deployment Already Has Them

Today we're shipping three new capabilities into Progress Agentic RAG: a Smart Agent, a native Microsoft Teams app, and a WordPress plugin. If you already run the platform, there's nothing to install or migrate, they arrive as configuration on the agentic knowledge foundation you're already running on. That's the real story here: a foundation is only as good as what it can absorb later, and this release is proof it absorbs a lot without a rebuild.

Smart agent for whoever's hand-writing this logic today

If you're the one deciding which source to query next, when to re-query, and how to stitch a multi-source answer together in application code, that job just got easier. That logic is now a capability of the platform, not a maintenance task your team owns forever.

The Smart Agent plans a question, breaks it into sub-questions, checks whatever sources are worth checking, and evaluates whether the answer is actually sufficient before it stops. Quality scoring runs at every step, so a weak link surfaces where it happens not three steps later as a confidently wrong answer.

Here's the big one: it doesn't stop at your indexed content. The Smart Agent also reasons across MCP-connected business applications — a CRM, an ERP, a ticketing system — pulling live records into the same answer alongside everything already in your knowledge base. One question can now combine an account's CRM history with an open support ticket and a contract that was never indexed anywhere, without a separate integration project for each

Microsoft Teams for whoever gets asked about usage, not just accuracy

If you approved the budget for an AI deployment, you already know the number that comes up in that conversation. It's usage, not retrieval quality.

The new Teams app, distributed through the Microsoft store, delivers the same indexed content, the same access controls, and the same source-cited answers your deployment already returns inside the tool your organization already has open. No bot framework, no custom auth flow, no new interface to learn.

WordPress plugin: the quiet proof

If you've been waiting to see whether this foundation holds up over time, this is the small, unglamorous evidence. WordPress content was already reachable through a crawl connector, so this isn't a new access point — it's a precision upgrade.

A native, source-side plugin keeps structure and metadata intact and content current, instead of stale behind a crawl. A foundation that only adds new sources is growing. One that also deepens the sources it already has is maturing.

Together, these three capabilities move on three different axes where an answer gets delivered, what kind of question the system can handle, and how precisely it reads the content it depends on. Every existing deployment already has all three.

Governance is the quiet risk behind every agentic AI rollout. The more autonomous a system gets — planning its own steps, deciding what to check next — the harder it becomes to explain afterward why it did what it did, and that's exactly where a promising pilot stalls out in a security or compliance review.

Source citations, quality scoring, and audit logs that ship on day one, not bolted on later, are what keeps systems thriving. Every step the Smart Agent takes traces back to where the answer came from, so a reviewer isn't being asked to just trust the outcome — they're able to check it.