Knowledge Graphs in Litigation Agents
summary
Company pipeline scrapes the web, extracts and schemas data into knowledge graphs, and runs multi-agent workflows with human-in-the-loop checks to find class-action leads far earlier than manual methods. Systems flag signals within 15 minutes to flag a lead and typically need about a month to confirm a viable case. Graph-backed retrieval composes lawyer-specific reports; empirical caution notes 95% per-agent → 77% across five steps for chained agents without heavy guardrails. Whyhow.ai is the deployed product.
AI-written summary · about
chapters (9)
- 0:00Introduction to graph RAG
- 2:21Defining graphs and relations
- 3:45Multi-agent system usage
- 4:59Challenges in AI agents
- 6:53Legal industry applications
- 9:36Knowledge graphs in discovery
- 11:59Case research and reporting
- 14:36Car fire case study
- 17:17Future of legal AI
related talks
14mKnowledge Graphs as the Backbone of AI-First Software Delivery
32mMulti-Agent Query Generation w/ LLMs on Complex Knowledge Graphs
23mKnowledge Graphs and LLMs in Action
15mGraphRAG-Powered AI Agents: Transforming Corporate Metadata into Intelligent Knowledge
21mAgentic GraphRAG: Autonomous Knowledge Graph Construction and Adaptive Retrieval
29mPowering Legal Research with Knowledge Graphs and Large Language Models
12mAgentic GraphRAG: Autonomous Knowledge Graph Construction and Adaptive Retrieval
28mAgentic GraphRAG: Multi-Agent Knowledge Graph