S&P debuts GenAI ‘Document Intelligence’ for Capital IQ

S&P Global will release a new generative AI-enabled tool as part of its Capital IQ Pro data workstation tomorrow, November 12. The offering is designed to help financial professionals save time and glean better insights when analyzing lengthy text-based documents.

The Document Intelligence tool is one result of a year-long and ongoing effort to integrate AI into Capital IQ Pro, which has also yielded the ChatIQ GenAI-enabled search interface for querying all data in Capital IQ Pro.

“We have a large corpus of textual information, such as 15 years of earnings call transcripts covering 11,000 companies, and investor presentations, news, and filings. So, how can we help our clients get insight out of all that textual information?” says Warren Breakstone, head of Capital IQ Solutions at S&P Global Market Intelligence.

Accessed via the “Documents” tab on the menu bar across the top of their screen, users can search for a company or topic and see all matching results across all types of documents—company financial results, filings, and news. But now, when a user clicks on any of the list of documents returned in their search, the system processes each document and displays it with additional, GenAI-generated information.

For example, to save the time required to scour a 50-page 10-Q filing, users can type questions, and the system provides a summary of the information they’re looking for. It also suggests what it thinks are the most useful questions to ask about specific companies and documents. Users can also compare that 10-Q to previous filings, and the system automatically highlights specific areas that have changed from one period to another—for example, changes in financial targets, or progress toward previously stated goals—and will summarize answers to users’ questions about any changes.

Users can apply similar comparison to earnings calls to see what has changed between announcements, and also to highlight other areas of interest, such as what other entities are mentioned during a call, how many times that company is mentioned, and how the number of mentions of that company has changed over time. Users can also click through from the mention of any company to view its profile in Capital IQ Pro, showing company information, stock price, news about the company, and other information.

For each document, such as an earnings call transcript, Document Intelligence also generates a side-by-side summary of positive and negative items mentioned, summaries by topic—such as discussions of strategy-related initiatives or product updates—and scores for sentiment and language complexity, along with comparisons of these scores to previous earnings calls as well as to earnings calls of the company’s peer group.

The ability to compare and analyze documents in this way relies on two layers of processing: one that takes place when the vendor first ingests a document, applying tags to each entity and its associated metadata. Then there is a second layer of processing that is performed instantaneously each time a user opens a document (about a second or two to add the AI-generated enhancements), which leverages AI technology derived from S&P’s 2018 acquisition of Kensho.

Breakstone says feedback from early adopters across the buy side, sell side, and corporate clients has been positive, with users saying it delivers clear time savings. In addition, he says, some firms that tested it have asked for the ability to run the same document analysis on their proprietary data, such as internal research or competitive analysis. That may involve having clients upload proprietary documents into Capital IQ Pro, or perhaps selling the GenAI Document Intelligence software on a standalone basis and is something the vendor is working on now, he adds.

“We approached Document Intelligence by thinking about the research and analysis workflow,” Breakstone says. “We’re not leading with technology to solve a problem; we’re leading with, ‘What’s the client’s problem?’ and then we work back to what data we have and what technologies can we embrace to bring that breadth and depth of data to bear for clients in many different relevant ways.”