Kensho NERD

Kensho NERD (Named Entity Recognition and Disambiguation) is an advanced AI-powered system that extracts and links entities—such as companies, people, places, and events—from unstructured text to structured knowledge bases like S&P Capital IQ and Wikimedia. It transforms raw documents into rich, searchable data, enabling deeper insights and smarter automation across industries.

Features

• Entity extraction from unstructured text
• Disambiguation using context-aware machine learning
• Linking to S&P Capital IQ and Wikimedia knowledge bases
• REST API for real-time and batch processing
• JSON-based annotations with entity metadata
• Confidence scoring for entity recognition and linking
• Support for aliases, abbreviations, and ambiguous terms
• Optimized for financial and business-related documents
• Scalable infrastructure for enterprise use
• Integration with other Kensho services like Link and Scribe

Capabilities

• Identify and classify entities in emails, reports, transcripts, and articles
• Link entities to structured profiles with financial, organizational, and biographical data
• Automate document annotation and enrichment workflows
• Enhance search, recommendation, and routing systems
• Detect sensitive information for redaction or compliance
• Support AI/ML model development with high-quality labeled data
• Handle complex entity relationships and corporate hierarchies
• Enable cross-domain insights through dual knowledge base linking

Benefits

• Transforms unstructured data into actionable intelligence
• Reduces manual effort in data tagging and document processing
• Improves accuracy and depth of financial and business analysis
• Enables faster decision-making with enriched context
• Enhances customer-facing applications with smarter content
• Supports compliance and risk management through entity tracking
• Scales across teams and departments with minimal setup
• Unlocks new revenue and research opportunities from existing data