TL;DR:

  • Moody’s has extended its Microsoft partnership, making its credit ratings, research, entity data and news available inside Microsoft 365 Copilot via a dedicated Moody’s agent built on the Model Context Protocol.
  • The move brings Moody’s “connected intelligence” — 600 million entities and 2 billion ownership links — into Copilot Chat, Researcher, Excel and related surfaces as a grounding data source, not just a plugin.
  • For UK financial-services firms, the pattern matters more than the announcement: MCP-distributed data partnerships are becoming the default AI enterprise architecture, echoing LSEG’s 150-plus MCP customer disclosure this week.

Moody’s and Microsoft have announced the next stage of a strategic partnership, embedding Moody’s credit intelligence directly into Microsoft AI products. The integration spans two delivery channels: a dedicated Moody’s agent in Microsoft 365 Copilot distributed through the Model Context Protocol (MCP), and Moody’s decision-grade intelligence as a grounding data source across Copilot Chat, the Researcher agent and Microsoft 365 Copilot in Excel.

What is actually being shipped

The MCP-distributed agent removes the custom-integration step enterprise customers previously had to build — shared Microsoft and Moody’s customers can activate Moody’s credit ratings, entity data, research and news directly inside Copilot’s existing workflow surfaces. Moody’s connected-intelligence layer spans a claimed 600 million entities and 2 billion ownership links, positioned as the context layer that makes AI output explainable and auditable for financial decisions. Moody’s CEO Rob Fauber framed the step as scaling access beyond specialists to “every person across an organisation who needs it”.

Why this matters in UK financial services

The announcement is the second major MCP-led integration disclosed this week. On Tuesday, the London Stock Exchange Group confirmed more than 150 customers now connect to or are integrating its own MCP server, channelling LSEG data into AI and cloud environments. The signal is consistent: MCP is becoming the default protocol for bringing regulated financial-services data into enterprise AI workflows, and UK banks and asset managers deploying Copilot will increasingly need to negotiate access to multiple data-vendor MCP endpoints rather than running bespoke integrations.

The compliance dimension is the other half of the story. Moody’s pitch on “explainable and auditable” output is directly aimed at the regulatory pressure UK financial firms are under — from the FCA’s AI Live Testing Sandbox, from the Bank of England’s CMORG cyber-resilience asks and from the upcoming joint FCA/PRA AI supervisory expectations. An AI that grounds answers in Moody’s rated data is an easier compliance conversation than one pulling from general-purpose web retrieval.

Looking forward

Expect the next six months to bring similar announcements from S&P, Fitch, Bloomberg and FactSet — every major financial-data vendor now has commercial pressure to ship an MCP-exposed layer or lose position in Copilot and Claude enterprise estates. UK procurement teams should ask their FS data vendors specifically about MCP roadmaps, pricing and data-residency terms before the next renewal cycle. The vendor-by-vendor licensing economics will define how much of a firm’s AI spend goes to Microsoft, Anthropic or OpenAI versus the content owners whose data makes the models useful.

Share this article