Kensho Blog S&P Global Launches Kensho LLM-ready API (beta), Making its Structured Data Accessible for…

S&P Global Launches Kensho LLM-ready API (beta), Making its Structured Data Accessible for Generative AINew solution enables customers to seamlessly integrate complex, high-priority S&P Global datasets into Generative AI models.Today, S&P Global announced the launch of a new solution in open beta that enables customers to access several high-priority S&P Global datasets for generative AI (GenAI) use cases. Kensho LLM-ready API integrates seamlessly with large language models (LLMs) like GPT, Gemini, or Claude, allowing customers to use natural language to query S&P Global’s tabular datasets. The API will help engineering, product, and business teams across financial services save significant time and resources by making company information, financial statements, historic market data, and global securities data a question away.Kensho LLM-ready API integrated into ChatGPTTraditionally, incorporating complex financial data into GenAI applications requires financial institutions to conduct arduous, time-consuming, and expensive data preprocessing, integration, and monitoring. As a result, many firms do not have the resources to start or maintain this process, which means they have not been able to fully take advantage of GenAI for their use cases. With the Kensho LLM-ready API, S&P Global has done this work already, so customers can add S&P Global’s rich, tabular data seamlessly and reliably into their LLM-enabled applications. Business users, such as analysts and traders, can write their queries into their institution’s chosen LLM and receive a trusted, timely answer. The solution draws on S&P Global’s financial domain expertise, which includes a sophisticated knowledge of the needs, use cases, and shorthand preferred by financial professionals. To support traceability and give customer confidence in the accuracy of the response, the solution provides the function calls or generated code as an audit trail.“As customers continue to integrate GenAI into their businesses, we want to make our data more easily and immediately accessible to the entire LLM ecosystem,” said Bhavesh Dayalji, Chief AI Officer at S&P Global. “We are providing the same high-quality, trusted data our customers rely on, now ready for our customer’s GenAI application of choice–be it their own internal proprietary LLM or a third-party application. We are expanding how and where customers can discover and leverage our offerings, with a vision to be the go-to partner for customers looking to capitalize on their investments in AI.”Asking natural language questions of complex structured data has historically been very difficult, due to high complexity of business and finance datasets, including a reliance on tabular data. As a result, without readily consumable data formats, financial institutions that want to harness GenAI are seeing delayed time to value. Kensho LLM-ready API, which has been used by a select group of customers in beta since April 2024, has already helped institutions speed up their GenAI adoption and workflows. The solution provides access to S&P Capital IQ Financials, Compustat® Financials, and Market Data, and partial access to Key Developments and GICRS, with plans to make additional datasets available throughout 2025.S&P Global is committed to getting more of its data “AI-ready” — which means it is trusted, formatted, and accessible to LLMs — allowing institutions to focus on leveraging AI for innovation and decision-making, instead of data preparation. It’s one of S&P Global’s many initiatives to help customers harness the full potential of GenAI, including providing the financial industry with robust LLM evaluation solutions through S&P AI Benchmarks and developing sophisticated GenAI interfaces into existing products, such as ChatAI on Platts Connect or ChatIQ in Capital IQ Pro.To learn more, find the solution on Marketplace or in the GPT store.S&P Global Launches Kensho LLM-ready API (beta), Making its Structured Data Accessible for… was originally published in Kensho Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.