Upload the room
Data rooms, Drive folders, financial models, PDFs, spreadsheets, decks, and contracts become a single source-linked corpus.
PE diligence intelligence · V2 now live
Build the deal brain once.
Run it across every diligence workflow.
Upload the data room. Provenance reads every document, maps the business, surfaces what is material, flags what is missing, catches what contradicts, and shows its work down to the source.
AI VDRDecision CockpitRisk AuditReasoning TraceIC Memo + DOCXAgent WorkflowsQ&ARequest List
Atlas Mobility · Deal BrainHold pending evidence
Coverage86%Financials, fleet, contracts, QoE
IC blockers4Revenue quality, capex, utilization
Material finding · CRITICAL
Adjusted EBITDA depends on utilization improvement that is not yet proven in source data.
Financial model → management deck → fleet schedule → risk auditSourceFincMod_Consol.xlsx
OntologyMetric → EBITDAQuality → Finding
OutputIC condition + management request
ExportIC Memo · DOCX ready
The problem
The data room is not the deal.
PE diligence is not a document search problem. It is a relationship problem. The material issue is usually the contradiction between two sources, the dependency across three companies, or the absence of a document everyone assumed existed.
Which revenue claim in the CIM conflicts with the customer files?
Seller story vs. source dataWhat obligation is buried in the contracts but missing from the model?
Hidden downside riskWhere are customer or regulatory exposures concentrated?
Risk invisible in one documentWhat evidence is missing from the data room that should be there?
Gaps analysts may not know to ask for219real VDR files tested
3,683source chunks processed
22risks surfaced
8IC memo sections
From VDR to conviction
Not another document chatbot.
A repeatable diligence system.
Provenance models the work an associate, VP, operating partner, and IC actually need: read the room, interrogate the business, and defend every conclusion with evidence.
Build the deal brain
Provenance maps the company: metrics, entities, obligations, risks, contradictions, coverage gaps, and source evidence.
Run diligence workflows
Financial, legal, operating, valuation, and IC workflows review the deal from every buyer angle.
Generate IC-ready outputs
Memo sections, risk audits, question lists, evidence packs, and decision traces your IC can verify.
Platform coverage · V2
The full diligence intelligence stack.
Live surfaces ship with full feature parity. Next and Roadmap items are under active build. Every live feature is wired to real API data — no static demo payloads.
Live
AI VDR
Full-corpus search, document intelligence, folder coverage, chunk-level source trails, and Not Analyzed visibility.
Live
Decision Cockpit
Proceed / hold / kill framing with confidence, blockers, conditions, priority asks, and evidence quality.
Live
Risk Audit
Severity, likelihood, audit status, IC impact, mitigation, and source quote for every material risk.
Live
Reasoning Trace
Follow a claim from source document to extraction, ontology class, diligence judgment, and IC output.
Live
IC Memo + DOCX
Partner-grade memo sections backed by source citations, valuation context, and request-list gaps. One-click DOCX export.
Live
Management Requests
Auto-generated diligence asks tied to missing evidence, contradictions, and unresolved blockers.
Live
Q&A / Ask
Grounded natural-language answers to partner questions — every response cites the source chunk, quote, and confidence score.
Live
Agent Workflows
6 diligence playbooks: financial, legal, operating, valuation, IC-ready, and full-stack. Critic-reviewed, trace-auditable runs.
Next
Datasite / Intralinks ingest
Native ingestion from major VDR platforms — files, folder structure, and permissions preserved.
Next
CapIQ / PitchBook comps
Pull public market comps and financials directly into the deal brain and valuation workflow.
Roadmap
MCP / API integration
Connect Provenance to firm workflows via Model Context Protocol or REST API. Bidirectional deal data.
Roadmap
Cross-deal intelligence
Prior-deal memory, sector pattern library, and firm knowledge hub surface relevant precedent across deals.
Why ontology
Most AI for finance is a vector store with a good prompt.
Provenance is built on Ontos — an enterprise ontology layer that models entities, contracts, obligations, exposures, assets, metrics, and time. Embeddings find similar text. Ontology answers questions a partner would ask.
Sample deal output
Give partners the answer, not another search box.
Provenance V2 — deal intelligence output · Atlas Mobility
// Processing: 219 diligence files · 33 not analyzed
// Build time: 47 seconds
DEAL BRAIN BUILT
Documents: 186 Chunks: 3,683 Risks: 22
ISSUE REGISTER — RANKED FOR IC
● CRITICAL Adjusted EBITDA depends on utilization improvement not yet proven in source data
● HIGH Revenue quality — run-rate claims unreconciled across CIM and financial model
● HIGH Fleet CapEx intensity not reflected in management projections
● MEDIUM Missing evidence — supplier concentration schedule absent from data room
OPEN DILIGENCE ASKS
22 follow-up questions generated for management and counsel
✓ IC memo generated // Decision Cockpit ready // DOCX export available
Feature parity · what's live vs what's next
Built toward the full finance AI workflow stack.
Parity on workflow surface. Dominance on trust surface. Every claim is traceable to an ontology fact. Every agent run is auditable. Live items are wired to production API. Next and Roadmap items are scoped for build.
Workflow agents
Diligence agents for financial, legal, operational, valuation, and IC workflows.
Deal intelligence layer
Connected VDR corpus, financial models, entities, risks, requests, and claims.
Document generation
IC memos, risk reports, diligence request lists, and evidence packs. DOCX export.
Knowledge graph
Enterprise ontology foundation for entities, obligations, assets, assessments, and findings.
Source-linked answers
Every material answer links back to documents, chunks, quotes, and confidence.
Agent reasoning trace
Full audit chain from source document → ontology fact → diligence judgment → IC output.
Enterprise controls
Private storage, tenant-aware access, auditability, and no public confidential deal data.
Native VDR platform ingest
Datasite, Intralinks, iDeals — direct API ingestion preserving folder structure and permissions.
Market data integration
CapIQ, PitchBook comps pulled directly into valuation workflow without copy-paste.
MCP / API for firm workflows
Connect to firm portals, CRM, and workflow tools via Model Context Protocol or REST.
Integrations and outputs
Meet the firm where the work already lives.
Datasite / Intralinks ingestOneDrive / SharePoint / Google DrivePitchBook / CapIQ / market dataCRM deal creation and syncDOCX / PPTX / PDF exportsMCP / API for firm workflows
Enterprise readiness
Your deal data stays private, auditable, and source-controlled.
- Private object storage and tenant-aware API architecture
- Claim-level citation validation for all material outputs
- Audit-ready trail from source document to IC recommendation
- Public demos use synthetic data only — confidential VDRs stay internal
- Model Context Protocol (MCP) and REST API for firm-controlled integration
For teams that need diligence they can defend