Summary

Upload docs → Tensorlake + GPT-5 → score firm-specific criteria; show clause excerpts & distribution vs. peers.

Problem Statement

Legal teams review hundreds of similar agreements but lack systematic ways to compare terms across documents. Important clauses like liability caps and termination rights vary widely, and outliers can create significant business risk if missed during review.

Details

Targets a long list of firm-defined checks (e.g., termination, indemnity, liability caps, MFN, non-compete scope). Produces per-dimension scores with linked clause excerpts and plots where each doc falls on a historical distribution to flag outliers. Built to speed reviews while keeping traceability to source text.

Images

What We Learned

Scoring agreements against firm-specific criteria catches risky outliers that manual review might miss. Showing where terms fall on historical distributions provides instant context for negotiation. Maintaining links to source clauses is critical for legal professional trust. Systematic analysis reveals which counterparties consistently push aggressive terms.