AI-Powered Legal Intelligence

Your documents already have the answers

Lumen reads every document in your case, extracts entities, maps relationships, and builds a queryable knowledge graph — automatically. Ask questions in plain English. Get sourced answers in seconds.

Explore a Real Case
Thousands of court filings, fully analyzed.
Knowledge graph built. Relationships mapped.
Ready to query.
Launch Demo →
Document review hasn't changed in decades
Until now.

The Old Way

  • Weeks of manual document reading
  • Relationships discovered by accident
  • Timelines reconstructed from memory
  • Critical evidence buried in attachments
  • New team members start from scratch
  • Keyword search misses what matters

With Lumen

  • Every document read and indexed in hours
  • Relationships mapped automatically across all filings
  • Timeline built from every dated reference
  • Semantic search finds meaning, not just words
  • New team members productive in minutes
  • Ask questions, get sourced answers
A knowledge graph that builds itself — then lets you explore it

Most legal AI tools search documents. Lumen understands them — extracting every person, organization, role, and relationship, then weaving them into a unified knowledge graph. Automatically.

Then it hands you the controls. Click any entity to see their full document footprint. Trace relationship paths between parties. Filter by connection type. Discover intermediaries buried across thousands of filings that no keyword search would ever surface.

  • Entities extracted, deduplicated, and linked across all documents
  • Relationships typed — employment, legal, financial, familial
  • Trace paths between any two entities in your case
  • Events placed on a unified timeline with source attribution
  • Every node and edge links back to the documents that prove it
Entity
Corp A
Person
Firm
Agency
Exec
Counsel
Filing
Witness
Dept
Bank
Auto-generated knowledge graph — click any node to explore
Two-stage AI that reasons over your case

Lumen doesn't just retrieve documents — it reasons across them. A fast planner analyzes your question and identifies relevant entities in the knowledge graph. Then a powerful synthesizer reads the evidence, cross-references across filings, and builds a sourced answer.

Hybrid retrieval combines knowledge graph traversal with semantic document search — so you get answers that understand both the relationships and the content.

  • Natural language questions — no query syntax to learn
  • Graph traversal + semantic search working together
  • Every claim cited with links to source documents
  • Pennies per query — built for daily use, not special occasions
Your Question
"What is the relationship between Company X and the regulatory filing from March?"
Stage 1 — Planner
Analyzes intent, identifies relevant entities in the knowledge graph, formulates retrieval strategy
Hybrid Retrieval
Knowledge graph traversal finds entity connections. Semantic search finds relevant passages. Both feed the synthesizer.
Stage 2 — Synthesizer
Cross-references evidence, resolves contradictions, produces a comprehensive answer with document-level citations
Six integrated tools, one unified case view
Everything works together. Search informs the graph. The graph informs the timeline. Every view links back to source documents.

Semantic Search

Search by meaning, not just keywords. Hybrid mode combines semantic understanding with exact matching to eliminate false negatives.

Entity Explorer

Every person, organization, and location — automatically extracted, deduplicated, and linked. See roles, relationships, and document frequency at a glance.

Relationship Graph

Interactive network visualization. Filter by relationship type, zoom into clusters, trace paths between any two parties.

Document Intelligence

AI-generated summaries for every filing. Browse by type, search across summaries, click through to the original.

Case Timeline

Every dated event extracted and placed chronologically. Filter by keyword, trace how events unfolded, link to sources.

AI Case Assistant

Ask questions in plain English. Get sourced answers that cross-reference the knowledge graph, documents, and timeline.

From raw filings to knowledge graph
Lumen's multi-stage AI pipeline processes your documents automatically.

Upload

Drop your case documents — PDFs, court filings, exhibits, depositions. Any volume.

Extract

AI reads every document, extracting entities, relationships, dates, and amounts.

Connect

Entities deduplicated and linked. Relationships categorized. Knowledge graph emerges.

Discover

Search, explore, query — find what you need in minutes, not weeks.

See what's hiding in your documents

Try the live demo with a real case. Then imagine what Lumen could do with yours.

Try the Demo
Or reach out at hello@cognivance.org