S&P Global’s March announcement of enhancements to its Capital IQ Pro platform covers a lot of ground: new datasets across fixed income, biopharma and private markets, the integration of two AI acquisitions, and expanded generative AI capabilities. Taken individually, each is a solid incremental update. But taken together, they point to something more structurally significant, a deliberate repositioning of Capital IQ Pro as an agent-native platform designed to be consumed by AI systems as much as by human analysts.
The centrepiece of that shift is S&P’s decision to expose Capital IQ Pro’s data and capabilities through Model Context Protocol (MCP) servers – the open standard that allows AI agents to interact with external tools and data sources. This is not a peripheral experiment. According to Warren Breakstone, Head of Data & Research at S&P Market Intelligence, it represents one of the two core pillars of the firm’s generative AI strategy.“A key component of our strategy is making the platform agent-friendly by unlocking it through MCP servers, so clients can access the same data and capabilities through various GenAI front-ends,” Breakstone explains to Market & Alt Data Insight. “Those MCP servers enable us to build agentic solutions ourselves and also enable our clients to build their own agentic solutions on top of Capital IQ Pro.”
The practical implication is that Capital IQ Pro is being opened up as data infrastructure rather than locked down as a proprietary terminal. Third-party AI agents – whether built by clients or by partners such as Anthropic – can query the platform, retrieve data, and present it within their own interfaces, while S&P retains control through a robust entitlements layer and direct contracting with clients.
That entitlements architecture is underpinned by a single access point that S&P has built in collaboration with Kensho, its in-house AI subsidiary acquired in 2018. “Working with Kensho, we’ve built out an access point called the Kensho Grounding Agent, which allows clients to access the full complement of our data through a single door,” says Breakstone.
Crucially, S&P’s long-established click-through-to-source capability – the ability to trace any data point back to its originating document or methodology – carries through into the agentic layer. “A client can access our data through Claude or other third parties and still have the ability to click through to the source documents and methodolgies used in Capital IQ Pro, ensuring confidence in the data and LLM,” Breakstone notes. For institutional users whose compliance and audit requirements demand full data provenance, this is a material consideration and one that distinguishes the approach from AI tools that surface answers without exposing their working.From Acquisition to Platform: The With Intelligence Test Case
The March release also provides a concrete illustration of how S&P operationalises acquired content and the speed at which it can do so. The firm acquired With Intelligence in Q4 2025, and by March 2026, the first tranche of its private markets content – more than 20,000 investment committee meeting packs, tender documents and LP filings – was live inside Capital IQ Pro’s Document Intelligence tool.
That velocity is enabled by the same Kensho pipeline that powers the Grounding Agent. Rather than manually normalising thousands of unstructured private markets documents, S&P runs them through automated entity-identification tooling that structures, tags and links the content to existing S&P reference data.
“That process identifies all the entities mentioned in each document and assigns discrete unique identifiers that map to S&P reference data,” Breakstone explains. “So we’ve effectively structured the content, tagged it, linked it, identified the entities, and brought it into the GenAI Document Intelligence solution.”
For data managers and quant researchers who have long grappled with the challenge of extracting structured, queryable insight from private markets documentation, the pipeline is arguably more significant than the documents themselves. It represents a repeatable mechanism that S&P can apply to future acquisitions and new content sets, a point that will not be lost on competitors evaluating their own build-versus-buy strategies.
The Enterprise Pricing Bet
Equally notable is how S&P intends clients to pay for all of this. In a market where modular, per-dataset pricing is the norm – and where entitlement complexity is a persistent operational headache for market data managers – Breakstone is explicit that S&P is taking a different approach.
“We minimise add-ons: Document Intelligence, for example, is part of the subscription. The With Intelligence investor documents I just described are part of the existing subscription,” he says.
The enterprise model – in which clients pay for Capital IQ Pro and can add unlimited users without incremental per-module charges – is designed to reduce friction and accelerate adoption, particularly for GenAI use cases where the value of the platform increases with the breadth of data accessible through it. It is a bet that bundling drives stickiness more effectively than monetising each new capability individually.
Competitive Landscape
S&P is not operating in a vacuum. Bloomberg, LSEG, FactSet and a growing cohort of AI-native challengers including AlphaSense/Tegus are all embedding generative AI into their research platforms. Regarding differentiation, Breakstone points to the combination of data breadth and AI capability rather than any single feature: the depth of the data estate, Kensho’s AI infrastructure, and the integration of both into a single platform that clients already rely on.
That is a reasonable strategic position, but it remains to be seen how it translates into observable competitive advantage as the rest of the market capitalises on agentic architecture and MCP adoption.
What Comes Next
Breakstone offers a preview of S&P’s June release, which he describes as significantly more expansive than the March announcement. “While this March release brings the first set of With Intelligence content into Capital IQ Pro, the June release is when the floodgates really open,” he says. That release will bring a much larger volume of With Intelligence data – news, commentary and differentiated content – into the platform, and make it available through S&P’s data feeds franchise for the first time.
The summer release will also introduce a refreshed version of Chat IQ, S&P’s ChatGPT-style conversational interface trained on Capital IQ Pro data, alongside the third version of Document Intelligence, which, for the first time, will allow clients to bring their own proprietary content into the platform and interrogate it using S&P’s GenAI alongside the firm’s own datasets.
“That almost completes the circle – clients using our GenAI, our vast dataset and their own data in combination,” says Breakstone.
The summer roadmap is worth watching closely. The combination of client-owned content, S&P reference data, and GenAI querying within a single platform – accessible both through a traditional interface and via AI agents through MCP – represents a significant step towards the kind of unified, agent-ready data infrastructure that the institutional market has been talking about but has rarely seen delivered in practice.
Subscribe to our newsletter
