S&P Global adds detailed broker forecasts to its AI Data Portal

Financial analysts can now bring the detailed broker forecasts behind market consensus into connected AI tools, rather than working only with headline estimates. S&P Global announced on October 15, 2026, that Visible Alpha data is available through its AI Data Portal, adding financial line items, business-segment forecasts and historical results to targeted research queries.

The immediate access method is S&P Global’s Deterministic Retrieval solution: direct, structured requests through an application programming interface, or API. It is intended for users who already know which data they need. S&P Global identifies earnings preparation, portfolio monitoring and equity research as uses for the newly connected information.

The detail beneath the consensus

Visible Alpha extracts and standardizes information from the full working models of more than 250 contributing sell-side brokers. According to S&P Global, that material covers more than 8,000 companies across over 170 industries. The integration therefore brings model-level detail into the portal, not simply a collection of top-line earnings predictions.

The available data falls into three groups, separating market expectations from reported results and allowing users to look below the company-wide total:

• Consensus estimates include driver-level metrics, contributor counts and estimate dispersion. Those fields show the inputs beneath an aggregate forecast, how many contributors are represented and how widely their estimates differ.
• Financial line items include standardized entries and historical actuals. This part of the offering supplies detailed financial records and past reported figures, distinct from the forward-looking estimates in the consensus data.
• Segment data includes forecasts and actuals at the business-segment and product level. Users can request information about those parts of a company rather than limiting their query to its combined results.

S&P Global’s stated rationale is that investors need both the assumptions behind a forecast and a view of analyst agreement. Peter Licursi, head of AI and Growth at Kensho Data & Platforms within S&P Global Market Intelligence, framed the addition around that distinction.

Investors and analysts want to see more than the headline number. They need the assumptions and the level of analyst agreement behind it,

Peter Licursi, Head of AI and Growth, Kensho Data & Platforms, S&P Global Market Intelligence

Access through connected AI applications

S&P Global says the integration is compatible with natural-language queries in third-party applications that support MCP, including GPT, Gemini, Claude and Copilot. The conversational interface and the retrieval method serve different roles: users can ask in ordinary language through a compatible application, while the portal provides targeted, structured access to the requested data.

Beyond preparing for earnings, S&P Global lists coverage initiation, comparable-company analysis, valuation and investor-relations benchmarking as supported workflows. These are uses the company identifies for the integration; the announcement’s emphasis is access to the underlying financial information through AI applications.

Broader retrieval remains a planned step

Visible Alpha support for Adaptive Retrieval is expected later in 2026, not included in the access available now. S&P Global describes that retrieval mode as designed for more complex, open-ended work, in which AI agents and language models retrieve and assemble information across multiple S&P Global datasets using natural-language queries.

Once that support arrives, the company says customers will be able to use either retrieval method or combine them. The practical dividing line is the task: a specific data request fits the current offering, while the planned extension is intended to handle questions requiring information assembled across datasets.