About WhyHow AI
Litigation intelligence for the $40B industry
WhyHow is a San Francisco-based litigation intelligence company that helps plaintiff firms identify emerging class action and mass-tort opportunities years before traditional detection methods. Using AI agents trained on top legal strategies, they leverage Exa search to surface weak signals - early indicators buried in niche articles, user forums, and regulatory filings - to identify potential litigation cases across the $40B industry.
The Challenge
Building litigation intelligence at scale
WhyHow AI needed to build a litigation intelligence engine that could identify early signals of legal issues across millions of web pages daily. They faced key obstacles:
- Brittle keyword search: Traditional search missed relevant documents and niche forums that never mentioned obvious legal terms but contained critical patterns for emergent drug reactions, latent automotive defects, and data-privacy abuses.
- Signal detection in noise: Early indicators of class action lawsuits were buried in isolated complaints, outlier spikes in warranty data, and terse footnotes in public filings across fragmented sources.
- Technical infrastructure gaps: Existing web search APIs and DIY stacks that re-rank Bing or Google results proved insufficient for context-rich recall for real-time legal monitoring.
The Solution
AI-powered litigation intelligence engine
WhyHow built an AI-powered litigation intelligence engine that scores millions of newly published pages daily for legal risk patterns, powered by Exa's search API with meaning based search instead of pure keyword methodologies.
Working with Exa's API, WhyHow created:
Real-time detection
Exa's search semantically finds pages across the web that mention emergent drug reactions, automotive defects, and data-privacy abuses by understanding meaning rather than relying on exact keyword matches. This enables WhyHow to process millions of pages daily and identify scattered complaints, warranty data spikes, and supplier disclosures that form patterns.
Legal reasoning integration
Platform that combines Exa's search results with LLM reasoning agents trained on legal strategies from top litigators to turn open-web text into actionable litigation intelligence.
What's under the hood?
Integrating Exa's Search + Crawling API
Qualities:
- Neural search
- Keyword-based search
- Relevant results
Integrate Exa Search
Key benefits of using Exa's Search + Crawling API include:
Semantic understanding
Find legal risk patterns that traditional keyword search misses in niche forums
Near real-time indexing
Fast indexing of newly published material critical for staying ahead of legal developments
Straightforward integration
Easy to integrate and fully customizable, allowing the team to leverage search & crawling in one API call
Business outcomes
30% higher accuracy and earlier case detection
WhyHow.AI deployed the platform and demonstrated the potential for semantic search to power applications far beyond conventional search functions.
30% higher accuracy
Exa returned correct source data 30% more often than other tools, especially on fresh articles critical for early detection.
Earlier case detection
Identify potential litigation opportunities years before traditional manual attorney research methods, turning scattered complaints into actionable legal intelligence.
Scalable monitoring
Process millions of pages daily across diverse sources, creating workflows impossible before LLM technology existed.
"In our internal benchmarks, Exa delivered accurate source data roughly 30% more often than any other tool we tried, especially on fresh articles. Its deep sub-page crawl also keeps our production agents from hallucinating."— Chia Jeng Yang, Co-Founder, WhyHow.AI
Looking ahead
Expanding monitoring and solving harder problems
Every large-scale tort begins with a pattern - scattered complaints, warranty data spikes, or supplier disclosures. As Exa continues building top-tier search infrastructure, WhyHow plans to expand monitoring across different practice areas and geographies, further shortening time-to-insight. The partnership highlights how when semantic search removes friction, LLM startups can tackle problems once dismissed as "too difficult to solve."