Kensho

Kensho is an AI engine under S&P Global. It integrates authoritative financial data into the workflows of large language models and intelligent agents, providing solutions such as deterministic retrieval, adaptive retrieval, document extraction, and entity linking, serving financial institutions and enterprises.

AI Legal Assistant#Financial Data#Entity Linking#Document Extraction

Updated: 2026-10-26

What is Kensho?

Kensho is an AI engine under S&P Global. It combines S&P Global's authoritative data with cutting-edge AI and engineering technologies to create reliable solutions for financial markets and enterprises. Unlike general-purpose large models, Kensho focuses on transforming complex and proprietary data into AI-usable intelligence, and connecting it to LLMs, intelligent agents and various AI applications, enabling customers to have more confidence in adopting next-generation technologies.

Who is it for?

Data and R&D teams at financial institutions that need to integrate S&P Global's trusted data into their own LLMs, agents or AI applications.
Sell-side and buy-side analysts who conduct company research, financial modeling, and cross-entity comparisons.
Quantitative and fundamental research teams that need earnings call intelligence and automatic integration of multi-source data.
Back-office data and compliance teams looking to batch-structured PDF reports and spreadsheet documents.
Enterprise customers with partnerships at S&P Global who wish to co-build customized AI solutions.

Key features

Deterministic Retrieval

Powered by the Kensho LLM-ready API, it accurately retrieves S&P Global data when requirements are clear, suitable for company research, financial analysis, earnings call intelligence and multi-entity comparison.

Adaptive Retrieval

Designed for complex and open-ended tasks, it dynamically routes across data sources, automatically generates queries, and compiles results, making it suitable for in-depth research, multi-step analysis, and agent-driven tasks.

Kensho Extract

Converts unstructured PDFs into AI-ready data, enabling fast and reliable document extraction across text, tables and charts, and reducing manual data entry costs.

Kensho Link

Connects tabular company data to S&P Global Company IDs and global identifiers, enabling large-scale intelligent entity matching to prevent name mismatches.

Core AI Capabilities

Structured and enriched proprietary data makes generative and agentic AI systems faster, more accurate, and more reliable, serving as the data foundation for upper-layer applications.

Kensho Labs Co-Creation

A team of AI experts and engineers collaborates hand-in-hand with selected S&P Global clients to innovate, developing prototypes and custom solutions for complex enterprise challenges beyond the standard roadmap.

How to use Kensho

First clarify your data needs: use Deterministic Retrieval (LLM-ready API) for precise data extraction, and Adaptive Retrieval for exploratory in-depth research.
use the two methods separately based on scenarios.
Use Kensho Extract to process backlogs of unstructured PDFs, then leverage Kensho Link to align extraction results with standard company IDs to avoid entity mismatches and duplicate work.
For complex enterprise-level challenges, you can collaborate through Kensho Labs co-creation, but this requires coordination via a S&P Global account manager and is not a self-service enabled product.
Compliance Tip: The scope of data usage is subject to the authorization and contractual terms of S&P Global.
For regulated data, you must comply with compliance and record-keeping requirements and must not make unauthorized calls.
Official capability descriptions may be adjusted with product iterations.
Before integration, check the current APIs, retrieval methods, and available data sources on the official solutions page.