VeritasGraph

The Connected Knowledge Graph

VeritasGraph links deal documents, participants, counterparties, assets, and contracts into a single queryable graph inside SpaceNexus — enabling multi-hop reasoning that goes beyond keyword search.

See the Diagram Request a Demo

VeritasGraph is the knowledge graph engine inside SpaceNexus that extracts entities from deal documents, maps typed relationships between them, and enables multi-hop reasoning queries across your data room. The VDR Document Pipeline feeds documents into the graph in real time, while entity extraction and relationship mapping power Generative Engine Optimization (GEO) — providing AI citability scoring, source attribution, and verifiable reasoning paths for LLM consumption.

Interactive Diagram

SpaceNexus Knowledge-Graph Explainer

How VeritasGraph connects deal documents, participants, counterparties, target assets, financials, and contracts into a single queryable graph through the VDR Document Pipeline.

Diagram: VeritasGraph knowledge graph connecting deal documents, participants, counterparties, target assets, financials, and contracts inside SpaceNexus. The VDR Document Pipeline serves as the ingestion layer feeding documents into the graph.

The knowledge graph diagram shows VeritasGraph at the center, connected to six entity types: Deal Documents (top), Deal Participants (left), Counterparties (right), Target Assets (bottom-left), Financials (bottom-right), and Contracts (bottom-center). The VDR Document Pipeline (bottom) feeds documents into the graph. Edge labels indicate deal relationships: filed by, negotiated between, acquired by, due diligence for, guaranteed by, and references.

Core Concepts

Connected Knowledge Graph

VeritasGraph turns your unstructured deal documents into a connected graph through three core capabilities.

Entity Extraction from Deal Documents

VeritasGraph parses your data room documents to extract deal-specific entities — buyers, sellers, advisors, target companies, facilities, revenue figures, EBITDA, and contractual clauses. Each entity is typed, scored, and linked to its source passage.

Relationship Mapping Across Deals

Entities are linked through typed deal relationships — 'negotiated between', 'acquired by', 'guaranteed by', 'references', 'located at'. Each edge carries metadata about its source document, page number, and extraction confidence.

Multi-Hop Reasoning for Due Diligence

Queries traverse multiple edges to answer complex deal questions. For example: 'Which contracts reference this facility?' requires hopping from a Facility node to Contracts via 'references', then to the signing Counterparty via 'executed by' — revealing connections keyword search misses.

Ingestion Layer

VDR Document Pipeline

The ingestion layer that feeds your data room documents into the knowledge graph through a five-step pipeline.

VDR Upload

Documents are uploaded to your SpaceNexus data room via the web portal, API, or bulk import.

Auto-Categorization

Documents are classified by type — financials, legal, correspondence, exhibits — using ML classifiers.

Entity Extraction

VeritasGraph extracts deal entities — participants, counterparties, amounts, dates, and clauses.

Graph Construction

Extracted entities and relationships are merged into the knowledge graph via VeritasGraph.

Incremental Updates

New documents trigger graph updates without full reprocessing — the graph stays current.

GEO Task

Generative Engine Optimization

How the knowledge graph ensures AI systems cite your deal data accurately and traceably.

AI Citability Scoring

Every answer generated from your data room includes a citability score — indicating how likely AI systems are to reference your content accurately.

Source Attribution on Every Answer

Every fact traced back to its source document with page number, paragraph, and extraction confidence — no black-box answers.

Verifiable Reasoning Paths

AI systems can trace the reasoning path through the graph — from question to answer — so you can audit how conclusions were reached.

Structured Data for LLMs

Knowledge graph triples are formatted as structured context that LLMs consume efficiently, reducing hallucination on deal-critical data.

llms.txt Integration

VeritasGraph can generate and maintain an llms.txt file that helps AI crawlers understand the graph structure and available deal data.

FAQ

Frequently Asked Questions

What is the SpaceNexus knowledge graph?

The SpaceNexus knowledge graph is a structured representation of all entities and relationships across your deals. It connects deal documents, participants, counterparties, target assets, financial statements, and contracts into a single queryable graph — so you can trace connections that span dozens of folders and hundreds of files.

How does VeritasGraph build the knowledge graph from my data room?

When documents are uploaded to your SpaceNexus data room, VeritasGraph ingests them through the VDR Document Pipeline. It extracts entities — deal participants, counterparties, monetary amounts, dates, legal clauses — and maps typed relationships between them. Each edge in the graph carries a confidence score and a link back to the source document and page.

What types of entities does VeritasGraph extract from deal documents?

VeritasGraph extracts deal participants (buyers, sellers, advisors, legal counsel), counterparties and legal entities, target assets and facilities, financial data (revenue, EBITDA, cap table entries), contractual obligations, and key dates. Each entity is typed, scored, and linked to its source passage.

How does multi-hop reasoning work in a due diligence context?

Multi-hop reasoning traverses multiple edges in the graph to answer complex questions. For example: 'Which contracts reference the Dallas facility?' requires hopping from the Facility node to Contracts via the 'references' edge, then to the signing Counterparty via 'executed by'. This reveals connections that keyword search alone would miss.

How does the VDR Document Pipeline feed the knowledge graph?

The VDR Document Pipeline is the ingestion layer that receives documents from your data room, auto-categorizes them (financials, legal, correspondence), triggers entity extraction, and feeds results into VeritasGraph for graph construction. New documents trigger incremental updates without full reprocessing.

What is GEO and how does it apply to my deal data?

Generative Engine Optimization (GEO) ensures that AI systems citing your data room content provide accurate, traceable answers. VeritasGraph powers GEO by scoring AI citability, attributing every fact to its source document, and providing verifiable reasoning paths through the knowledge graph.

Ready to see your deal data as a connected graph?

Upload your data room documents and let VeritasGraph build the knowledge graph automatically.

Request a demo →See pricing

Related Resources

Continue exploring

Feature