Hudson Labs
Hudson Labs provides an AI co-analyst platform for institutional US public markets research, extracting restatement-adjusted financial figures directly from unstructured SEC filings and other source documents. The platform models company guidance, detects fraud risk through proprietary forensic scoring, and reduces AI hallucinations by returning "N/A" rather than inventing data. It covers 10,000+ US public companies with 250M+ financial data points and is trusted by firms managing over $600B in assets.
- Data & Analytics
- Artificial Intelligence
- Financial Technology
- AI Agents
- Software Only
Funding
Funding not disclosed
Founders
Founder details are not available yet.
Product
Problem
Institutional investors analyzing US public companies face significant challenges extracting accurate, multi-period financial data from unstructured SEC filings and other source documents. General-purpose AI tools hallucinate figures, lack domain-aware financial logic, and cannot reliably interpret forward-looking statements, leading to costly errors and missed signals in investment research.
Solution
Hudson Labs provides an AI co-analyst platform purpose-built for institutional US public markets research. The platform uses proprietary retrieval systems, custom AI models, and agentic source selection to extract restatement-adjusted financial figures directly from unstructured documents, ensuring numeric accuracy. It models company guidance to identify subtle cues about future performance and offers AI-powered text search that combines natural language questions with comprehensive fundamental data. The Co-Analyst reduces hallucinations through confidence-based decoding and rigorous post-processing, returning "N/A" rather than inventing results when information is unavailable. The platform also includes forensic risk assessment that predicts the likelihood of fraud using unstructured data points like related party, accounting, and off-balance sheet risk. Trusted by firms managing over $600B in assets, the platform delivers accuracy, transparency, and efficiency for high-stakes investment workflows.
Target Audience
Primary customers are institutional investment firms, Fortune 100 companies, and top law firms conducting US public markets research, including analysts, portfolio managers, and legal professionals who need accurate financial data and fraud detection capabilities.
Features
- Restatement-adjusted multi-period financial figures extracted directly from unstructured source documents
- Guidance modeling that identifies subtle forward-looking cues about company performance, including hedged language
- AI-powered text search combining natural language queries with comprehensive fundamental data
- Proprietary retrieval, custom AI models, and agentic source selection designed for financial accuracy
- Confidence-based decoding and post-processing to reduce hallucinations, returning "N/A" when data is unavailable
- Red flag modeling and ranking for detecting potential fraud, going concern issues, and ineffective controls, with high-risk names 3x more likely to be sued for fraud
- Coverage of 10,000+ US public companies with 250M+ financial data points and 99% financial accuracy
- Co-Analyst Agents for automating recurring research workflows, plus Model Context Protocol (MCP) integration with Claude