LegalTechnical Co-Founder

Clause: AI Contract Review Platform

Built an AI-powered legal platform for contract drafting, review, and semantic search. Deployed as a Word add-in enabling lawyers to work with AI assistance directly in their existing workflow.

Visit clause.so

Key Impact

  • 14 organizations within 6 months
  • Natural language search across 1000s of ingested documents
  • Playbook review automatically aligns contracts to organizational policies
  • Productized from successful Augmented AI Labs client work

The Problem

Law firms and legal departments spend enormous amounts of time on repetitive contract work. Lawyers draft similar contracts over and over, manually review documents against organizational playbooks, and struggle to find relevant precedents across thousands of historical contracts.

Existing legal tech solutions either required lawyers to abandon their familiar tools (Word) or offered surface-level AI features that didn't integrate into actual workflows.

The Solution

I built Clause as a Word add-in that brings AI capabilities directly into the lawyer's existing workflow. No context switching, no new interfaces to learn. Just AI assistance where they already work.

Core Features

AI-Assisted Drafting & Redlining

Lawyers can draft and redline contracts with AI assistance directly in Word. The system understands legal context and provides intelligent suggestions that align with the firm's style and precedents.

Playbook Review

The playbook review feature automatically analyzes contracts against organizational policies. Instead of manually checking each clause against a 50-page playbook, lawyers get instant identification of non-compliant terms with suggested revisions.

Contract Explorer (RAG)

Natural language search across the entire contract repository. Lawyers can ask questions like "Show me indemnification clauses from our SaaS agreements with Fortune 500 companies" and get relevant results from thousands of ingested documents.

Technical Architecture

The system uses a RAG (Retrieval-Augmented Generation) architecture with Azure AI Search for semantic search capabilities. Documents are processed and chunked for optimal retrieval, with embeddings stored for similarity search.

I architected the system to be model-agnostic, supporting OpenAI GPT-4, Claude Sonnet, and Gemini. This flexibility lets us optimize for different use cases and provides resilience against API issues.

Key Technical Decisions:

  • Word Add-in Architecture: Built as an Office Add-in for seamless integration with existing workflows
  • Multi-model Support: Abstracted LLM layer allows switching between providers based on task requirements
  • Hybrid Search: Combined keyword and semantic search for better retrieval accuracy
  • Enterprise Auth: Full Microsoft and Google SSO integration for enterprise deployment

Team & Process

I led 2 offshore engineers through the full deployment lifecycle, from requirements gathering with law firms to customer feedback incorporation. We shipped iteratively, getting features into users' hands quickly and refining based on real usage patterns.

Results

Within 6 months of launch, Clause had been adopted by 14 organizations. The product was successfully productized from earlier client work at Augmented AI Labs, proving the market demand for AI-native legal tools that integrate with existing workflows.

Technologies Used

ReactFlaskAzure AI SearchAzure Blob StorageAzure App InsightsOpenAI GPT-4Claude SonnetGeminiMicrosoft AuthGoogle Auth

Facing a similar challenge?

I build AI solutions like this for companies ready to automate manual processes or unlock insights from their data. Whether you need an off-the-shelf tool configured or a custom system built from scratch, I can help.

Free consultation. I'll assess your situation and give you an honest recommendation.