If you have started using an AI assistant for IR work, you probably know both where it helps and where it stops. It is quick with a first draft of an earnings-call Q&A or a summary of a sell-side note. But it does not know your shareholder base. Ask about last quarter's 13F movement, the notes from your last meeting with a top-25 holder, or which underweight funds to target next, and a general-purpose assistant has nothing authoritative to work from.

What if it could work directly from your own investor data — live, permissioned, and accurate — without you ever leaving the conversation? S&P Global's Capital Access Model Context Protocol (MCP) offers precisely this capability.

What is the Capital Access MCP?

Model Context Protocol (MCP) is an open standard that acts as a bridge, securely connecting Artificial Intelligence (AI) assistants, such as Claude, ChatGPT, and Copilot, to systems where data resides. Instead of copying information into a prompt, an MCP allows an AI assistant to securely request live data and perform actions in a source system on demand. When you ask a question in natural language, the AI assistant chooses the most appropriate "tool" from the MCP, supplies the necessary parameters, and returns an answer grounded in the provider’s authoritative data.

The Capital Access MCP specifically gives clients conversational access to their Capital Access data, including ownership, institutions, contacts, and CRM activity, without needing to leave their AI assistant. This integration combines authoritative ownership data with engagement and interaction records, enabling users to seamlessly transition from queries about stock ownership to discussions about past meetings and topics.

Practical IR Uses for MCP

The Capital Access MCP connects our trusted, authoritative databases directly with your AI assistant. By bringing ownership, targeting, and CRM data into one conversational space, you can easily handle key tasks that used to require jumping between different systems:

• Query your shareholder base: You can instantly pull detailed ownership breakdowns, look at historical trends, or filter by specific investor details. This makes it easy to track who is buying or selling your stock, and spot any early, unusual position changes.
• Find and organize new investor targets: Instead of spending hours exporting spreadsheets to find underweight investors or active managers, you can just ask the assistant to find them for you. You can build, filter, and save qualified target lists directly to your profile.
• Prepare for your next meeting: The assistant can quickly scan your meeting notes and activity logs to summarize what you discussed during past interactions, helping you walk into your next investor meeting fully prepared.
• Plan smarter roadshows and events: You can ask the assistant to analyze your target list locations and suggest the most efficient cities to host a non-deal roadshow, complete with a list of local prospects.
• Build board reports in minutes: Instead of manually compiling statistics from different tools, you can ask the assistant to draft a quarterly update covering your top buyers and sellers, recent meeting counts, and progress against your targeting goals.
• Direct access for executives: Because access is available to all Capital Access users, the CEO or CFO can ask their own questions rather than routing every request through IR. Ahead of a meeting they can pull an investor's latest ownership position, your recent meeting history with the firm, and the notes from the last conversation — in seconds, in the assistant they already use.

Beyond Capital Access: One Conversation, Many Data Sources

Because MCP is an open standard, the Capital Access connection is rarely the only one your assistant is using. Add a Capital IQ Pro connection and a single request can pull together what used to be four separate exports: current ownership and position changes, consensus estimates and revisions, share price and peer performance, and summaries of recent earnings calls. Board reporting that once meant hours or days of assembling slides becomes a draft your team reviews and refines in minutes.

The same logic applies to your own systems. Connecting internal sources like SharePoint sites or folders, Outlook, or shared drives lets the assistant work from your prior board decks, IR calendars, and correspondence alongside external market data. The answer reflects both what the market is doing and what your team already knows, with no one copying figures between windows.

Ultimately, this lets you transition smoothly from asking "who owns our stock?" to "when did we last meet with them and what did they say?" It takes the heavy lifting out of pulling data, so you can spend less time on spreadsheets and more time building relationships. And as you connect more sources, the same conversation keeps getting more useful.