AI is rapidly becoming part of everyday financial workflows deeply embedded in the tools we all use. In investment and risk decisions, AI must be grounded in content that is precise, structured, and governed: trusted data you can compute with, explain and audit.

That is the rationale behind LSEG's approach. Through the Model Context Protocol (MCP), LSEG's licensed, trusted financial content, deterministic analytics, and semantic depth can now be accessed in Excel and Powerpoint via Anthropic's Claude. It is a ‘meaning layer’ that is directly in the analyst’s everyday toolkit, ready to support decisions.

Where deterministic meets probabilistic
The most important architectural decision in AI-powered finance is not which model to use, it is where to draw the line between what should be computed and what should be inferred.

Large language models are powerful at probabilistic tasks: synthesising narratives from earnings calls, identifying themes across a sector, and connecting macro signals to market moves. But many financial outputs are not a matter of interpretation. For example, normalised free cash flow, consensus estimates, yield curve analytics, spread duration, or portfolio risk contributions require deterministic computation with defined methodologies. When those numbers are wrong, the consequences show up in basis points and P&L errors, legal risk, and regulatory scrutiny.

This is why LSEG is so foundational in the AI workflow. LSEG’s trusted content does not simply give an AI model ‘more information’ to reason over, it provides critical deterministic anchors: standardised fundamentals, calibrated macro series, validated consensus estimates, and validated analytical outputs from products such as Yield Book.

Claude can then reason around those anchors, contextualising fundamentals against macro trends, comparing analytics to consensus, linking Reuters News to instruments and exposures, while the quantitative foundation remains exact, consistent, and auditable.

Semantic depth: teaching AI what finance actually means
Scale matters. LSEG brings extensive breadth across company fundamentals, estimates, macroeconomic data, cross-asset analytics and Reuters News, with new content continuing to be added. What makes Claude work reliably in finance is LSEG’s unmatched proprietary data - the semantic and ontological architecture underneath.

Every entity, instrument, and datapoint in LSEG’s data sets exist within a structured web of relationships: taxonomies that classify instruments across asset classes and jurisdictions; standardised definitions that support like-for-like comparisons across accounting regimes; identifiers that resolve ambiguity; and cross-referencing that links an issuer to its securities, peers, filings, and estimates.

This semantic depth enables an AI agent to decompose a question such as: ‘Show me investment-grade European corporate issuers with rising free cash flow and tightening CDS spreads’ into precise, executable requests, not probabilistic guesses stitched together from inconsistent sources. 

In other words: LSEG’s trusted data ensures AI is not only fluent, but financially correct.

Why MCP makes this practical and turnkey
MCP, originally developed by Anthropic and now governed as an open standard, is what makes this integration much easier. 

MCP is much simpler: standardised connectivity with security and licensing governance preserved.

For users, the experience is immediate. In Claude's Excel and Powerpoint experiences, LSEG customers can use natural language to retrieve and structure LSEG’s trusted financial content without the need for engineering or coding. In other words – it enables customers of LSEG to use more of its trusted content.

What this unlocks for LSEG clients
This combination of deterministic analytics, semantic depth, and new delivery reshapes how work gets done:

• Investment banking: peer screening and comps across jurisdictions become faster because standardised fundamentals and estimates arrive consistent and comparable, freeing time for judgement and advice.
• Wealth management: advisers can move across equities and fixed income with coherent analytics, modelling allocation changes with validated data and news context in the same environment.
• Quant research and trading: teams can explore and prototype datasets conversationally in Excel before production pipelines, accelerating iteration while preserving methodological rigour.

Across all these roles LSEG reduces the gap between accessing data and using it for actionable insight. LSEG customers with access to Claude’s products can benefit from deep integration today. 

LSEG Everywhere, in action
LSEG’s collaboration with Anthropic, including direct MCP connectivity, exemplifies this ecosystem approach: extending LSEG's reach into the AI layer while keeping the same licensing and governance principles intact.

The result is strategic flexibility for institutions. AI platform choice becomes a workflow decision, with LSEG's trusted data available natively, governed, and ready.

That is the ‘meaning layer’: precise enough to compute with, structured enough to reason over, and governed enough to act on, now in Excel and Powerpoint.

Read the original article here.