S&P Global Unveils Adaptive Retrieval for AI-Driven Financial Workflows

The new service functions alongside the company's existing Deterministic Retrieval method within a unified interface, the S&P Global AI Data Portal. While the deterministic approach relies on the Kensho LLM-ready API for precise, structured inquiries into specific company records or earnings transcripts, the adaptive layer is designed for complex, multi-step research. By enabling agents to synthesize data from disparate sources, the tool aims to reduce the engineering burden of validating and connecting datasets.

Sally Moore, Chief Client Officer at S&P Global, notes that the integration addresses the transition from human-led tasks to independent agentic systems. The company is positioning its data as an auditable, cited foundation for AI development, attempting to remove the friction of manual data preparation. According to Bhavesh Dayalji, who heads Kensho Data & Intelligence, the firm intends to expand this further by embedding its information directly into customer workflows through AI-native plugins and visualization tools, ensuring that proprietary intelligence remains at the core of evolving multi-agent architectures.