Contract Review & Risk Scoring

Part of the worlds-biggest-software-project initiative.

An open-source, AI-native contract review engine that scores risk against your own playbook, with explainable rationale and pluggable LLM backends.

Contract Review & Risk Scoring is a candidate open-source project for AI-powered contract analysis, risk flagging, and playbook enforcement. It is aimed at in-house legal teams, law firms, M&A and due diligence teams, procurement and compliance officers who need to review contracts faster without the lock-in or cost of enterprise CLM suites. The project addresses the gap between expensive proprietary tools (Ironclad, Kira, Luminance, Icertis) and lightweight assistants (Spellbook, LegalSifter) by offering a self-hostable engine grounded in open clause taxonomies.

Why Contract Review & Risk Scoring?

The CLM market is dominated by proprietary vendors with high switching costs; enterprise platforms such as Icertis run $150K–$1M+/year and Kira and Luminance start at $25K–$30K/year, placing structured contract review out of reach for most mid-market and public-sector organisations.
Existing AI contract review tools flag deviations but struggle to produce explainable, jurisdiction-specific risk scores tied to the customer's own negotiation playbook.
No leading commercial tool natively maps obligations across master agreements, SOWs, and amendments, which leads to undetected conflicts and commercial disputes.
No current tool monitors regulatory feeds and automatically identifies contracts containing affected provisions, leaving in-house teams to run weeks-long manual audits whenever a regulation changes.
No established open-source contract review engine with a structured clause taxonomy and open playbook format exists as of 2026, despite open standards such as SALI LMSS and IACCM frameworks being available as building blocks.

Key Features

Clause Extraction & Repository

Clause extraction from uploaded contracts (PDF, DOCX, scanned) using a pluggable LLM backend.
At least 50 pre-defined standard clause types out of the box.
Contract repository with full-text search, version history, and metadata tagging.
Custom clause type training from user-labelled examples via a few-shot or fine-tuning pipeline.
Export of risk summary reports as PDF and CSV for legal team or board review.

Playbook-Grounded Risk Scoring

Configurable playbook with preferred, acceptable, and fallback positions per clause type.
Risk-tier assignment per clause type and aggregate contract risk scores.
Plain-language rationale for every flagged clause, not just a binary flag.
Jurisdiction-aware risk calibration that adjusts scores based on governing law selection.

Multi-Document Obligation Graph

Linking of obligations, payment terms, termination rights, and renewal triggers across MSAs, SOWs, and amendments.
Cross-document conflict detection, alerting when a downstream document contradicts the master agreement.
Predictive renewal and obligation performance alerting derived from extracted dates and milestones.

Workflow, Approvals & E-Signature

Approval workflow with configurable routing, email notification, and a complete audit trail.
E-signature integration via DocuSign and Adobe Sign APIs.
Self-serve counterparty collaboration portal for negotiation without email attachments.

Regulatory & Adversarial Intelligence

Regulatory clause update monitoring, alerting when contracts contain provisions affected by specified new regulations.
Adversarial clause detection that flags clauses statistically correlated with litigation outcomes.
Natural-language contract Q&A, for example "what is the liability cap in this agreement?".
Benchmark comparison showing how a contract's risk profile compares to the organisation's historical portfolio.

AI-Native Advantage

Where incumbents flag deviations, this project produces explainable, jurisdiction-specific risk scores anchored to the customer's own playbook with plain-English rationale rather than opaque flags. It builds a knowledge graph of obligations, payment terms, and termination rights across master agreements, SOWs, and amendments, surfacing conflicts that no current tool reliably catches. AI models trained on litigation outcomes can identify provisions that statistically correlate with disputes and cost overruns, enabling predictive risk scoring beyond simple playbook compliance. A regulatory feed monitor automatically identifies affected contracts and suggests remediation language, compressing weeks-long compliance audits into hours.

Tech Stack & Deployment

The system is designed as a self-hostable open-source engine with pluggable LLM backends, allowing organisations to choose between hosted models and on-premise inference for sensitive matters. Clause taxonomy and playbook formats build on open standards including the SALI Alliance LMSS and IACCM / World Commerce & Contracting frameworks, and account for UNCITRAL, eIDAS, GDPR Article 28, and FASB ASC 606 / IFRS 15 obligations. Integration points target Microsoft Word and Google Docs for drafting, iManage, NetDocuments, and SharePoint for document storage, and DocuSign and Adobe Sign for execution. A REST API and webhooks support custom integrations into Salesforce, Workday, NetSuite, ServiceNow, Slack, and Teams.

Market Context

The contract lifecycle management software market is estimated at $3.0–$3.39 billion in 2025–2026 and growing at 11–14% CAGR through 2031–2034 (Mordor Intelligence; Fortune Business Insights; Precedence Research), with North America accounting for around 45% of the global total. Pricing for incumbents ranges from $99–$199/user/month for Spellbook, $25K–$50K/year for Kira and Luminance, and $150K–$1M+/year for Icertis enterprise deployments. Primary buyers are in-house legal teams, M&A and due diligence teams, large enterprise procurement and legal organisations, transactional law firms, and compliance officers.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.