In August we had our largest product launch ever around two themes: governing how research enters AI systems, and putting AI directly into the workflows research teams and their clients already use. Across those two themes we introduced several new capabilities. Two of them launched in beta.
Today, two of those capabilities take the next step.
Portal AI Search & Chat and Portal Connect MCP are generally available today.
What we learned in beta
Beta periods do more than test software. They show you how people actually behave when a new capability is in their hands. This beta revealed two things.
The first: Clients engage with research differently when AI is in the room. They do not search for a document. They ask a question, they follow up with another one, they explore a theme for twenty minutes without ever leaving the portal. The engagement data that came back, what we call the Three Cs of Context, Citation, and Concepts, gave sell-side research teams visibility into client intent that email open rates and page views have never come close to providing.
The second: The friction of context switching is more expensive than anyone had put a number on. Internal analysts, bankers, and sales teams who needed to access firm research inside their AI workflows had built workarounds. Opening the Portal UI, copying a link, pasting it somewhere else. Small steps, repeated constantly, across an entire organization. Portal Connect MCP eliminated that.
We spent the beta period watching, listening, and tightening – and both products are better for it.
Portal AI Search & Chat
Your clients are already using AI to search and summarize research. Most of that is happening outside your portal, outside your entitlements, and without any visibility back to you.
Portal AI Search & Chat changes that. AI-powered search and contextual chat, built directly into the portal your clients already use. Natural language questions with instant, entitlement-aware answers. Contextual follow-up without leaving the tool. And every interaction, every query, every citation, every concept explored, comes back to your team as a signal.
Not only does this enhance the overall client experience, but it also surfaces a new kind of intelligence about what your clients actually care about, in real time.
Note: Models are never trained on client data. Full audit logging, guardrail monitoring, and configurable disclaimers are on by default.
Portal Connect MCP
Your firm's research is one of its most valuable assets. Yet for most internal teams, accessing it means opening a separate application, breaking out of the workflow they are already in, and manually transferring what they find.
Portal Connect MCP fixes that. One governed MCP server connects your firm's full research corpus to any AI platform your internal teams already use. Investment banking, strategy, research sales - they get the research they need, inside the tools they are already working in, without a portal login and without a context switch.
Access is controlled by the firm. Every query is attributed, and every interaction is logged. And for the first time, research leadership has visibility into how the firm's own intellectual capital is being used internally, not just by clients.
What comes next
Both products are now in full production and available to all eligible clients. If your firm expressed interest during the Summer 2026 launch or has been waiting for the beta to close, the door is open.
Reach out to your BlueMatrix relationship manager or contact us at sales@bluematrix.com.