Kensho AI

Kensho AI

Kensho AI is S&P Global’s AI innovation hub, turning messy documents, earnings calls and news into clean, machine-readable data for finance, legal and commercial teams. Built on advanced NLP and machine learning, its APIs and ready-to-use models help analysts find numbers, entities and insights faster and with full audit trails.
Rating:
5
Visit Website
AI data extractionfinancial AI analyticsNLP named entity recognitionbusiness audio transcriptionintelligent document processingKensho ScribeKensho NERDconvert unstructured data to structured

Features of Kensho AI

Kensho Scribe – finance-grade speech-to-text for earnings calls and meetings, with real-time streaming and built-in jargon recognition.
Kensho Extract – batch-pull tables, text and key-value pairs from PDFs and image files, outputting machine-ready JSON or CSV.
Kensho NERD – detect and disambiguate companies, people, tickers and link them to S&P Global identifiers in a single call.
Kensho Link – map internal spreadsheets to S&P Global databases (Capital IQ, Compustat) for instant data cleansing and enrichment.
Kensho Classify – no-code AutoML tool to train custom document classifiers for filings, research, ESG reports, etc.
Grounding – LangGraph-powered multi-agent framework that stitches fragmented financial data sources into one searchable graph.
LLM-ready API (Beta) – turn natural-language questions into traceable, reproducible function calls for seamless LLM integration.

Use Cases of Kensho AI

Equity analysts upload earnings-call recordings and receive time-stamped transcripts plus auto-extracted KPIs within minutes.
Compliance teams scan thousands of contracts to surface clauses, parties and obligations, cutting review time by 70 %.
Macro researchers feed news, broker notes and filings into NERD to map event-driven relationships between companies and securities.
Asset managers reconcile internal holdings with external identifiers (CIQ ID, ISIN, CUSIP) via Kensho Link for portfolio risk checks.
Data engineers extract balance-sheet tables from 500-page PDF annual reports and pipe them straight into Snowflake or Excel.
Academics transcribe interview archives and auto-tag entities to build citation networks for empirical studies.
IT departments auto-label internal memos, decks and research to power enterprise search and knowledge-base tagging.

FAQ about Kensho AI

QWhat is Kensho AI?

Kensho AI is S&P Global’s artificial-intelligence division that offers NLP and machine-learning APIs to convert unstructured text and audio into structured, actionable data for financial and commercial use cases.

QWho are the main users of Kensho AI?

Financial analysts, portfolio managers, compliance officers, data scientists, market researchers and legal professionals who need to process large volumes of documents or audio files.

QHow do I start using Kensho AI products?

Request a free trial on kensho.com—you’ll receive API keys or GUI access with limited credits for Scribe (audio) and NERD (entity extraction). Enterprise pricing is available through S&P Global sales.

QHow accurate is Kensho Scribe?

Scribe is trained on finance and business vocabulary and optimized for accented English, multiple speakers and numeric entities. Optional human-in-the-loop review can push word-error rates below 3 %.

QHow does Kensho AI protect my data?

As an S&P Global company, Kensho follows strict encryption, SOC-2 controls and industry-specific compliance. Details are covered in the official privacy policy and data-processing addendum.

QIs Kensho AI free?

Core products come with a no-cost trial tier. Full-volume commercial use is licensed; contact sales for pricing tailored to your data volume and deployment model.

QHow does Kensho connect to S&P Global data?

NERD and Link automatically match extracted entities or spreadsheet rows to S&P Global identifiers (CIQ ID, Compustat GVKEY, etc.), giving instant access to fundamentals, estimates and ownership data.

QCan Kensho AI integrate with my existing systems?

Yes—REST & GraphQL APIs, batch file drop, WebSocket streaming and the new LLM-ready API plug into Snowflake, AWS, Azure, on-prem workflows or custom Python/R scripts.