The new service pairs natural-language access for AI agents with the existing Deterministic Retrieval, expanding how customers use licensed S&P Global data across agentic and structured workflows.
S&P Global said it has launched Adaptive Retrieval, a service that lets AI agents and large language models use natural-language queries to access and assemble licensed S&P Global data. The company is offering it alongside Deterministic Retrieval through a single product, the S&P Global AI Data Portal, positioning itself as the first to provide customers with both retrieval approaches in one solution. Adaptive Retrieval is designed for broader, multi-step tasks by allowing AI systems to pull from multiple data sources at once and handle requests spanning several datasets. Deterministic Retrieval, built on the Kensho LLM-ready API and available since 2025, is aimed at more structured, API-driven use cases such as researching a specific company or analyzing earnings call transcripts. S&P Global said customers can use either method or combine both depending on their AI architecture and workflow needs. The launch reflects a wider push by financial-data providers to make enterprise information easier for AI systems to consume while preserving citation, verification and auditability. S&P Global linked the rollout to the recent evolution of its Market Intelligence operating model and to the newly formed Kensho Data Platforms vertical, which it said is focused on building more AI-native interfaces and workflow tools.