Clause-level risk scoring
LiveRed, amber, or green for every clause, with an explanation and confidence score.
- ✓Per-clause red/amber/green tags
- ✓Plain-English explanations on every verdict
- ✓Confidence labels (low/medium/high)
- ✓Statute references where relevant
What it is.
Every clause in your contract gets a per-clause risk verdict: red (block / renegotiate), amber (worth a second look), or green (standard). Each verdict includes a plain-English explanation, a confidence score (low / medium / high), and a reference to the relevant statute or precedent if applicable.
Risk scoring runs over the parsed clauses produced by the scanner. The scoring model is trained across a broad range of commercial contract types.
Three steps.
End to end.
Document parses into clauses. Risk model runs over each clause.
Red/amber/green tag with a sentence-level explanation in plain English.
Click a clause to see the reasoning, the relevant statute, and the fix suggestion.
What you get.
- ✓Per-clause red/amber/green tags
- ✓Plain-English explanations on every verdict
- ✓Confidence labels (low/medium/high)
- ✓Statute references where relevant
- ✓Filter by risk level for triage on long contracts
Quick answers.
Override it. Overrides are logged and used to tune the model for your workspace (opt-in).
It catches a subset — missing clauses, contradictions, definitional gaps. Implied tax or compliance risks fall to Tax AI and Compliance AI respectively.
More in Review & Risk Analysis.
Three-tier extraction cascade so PDF, DOCX, and scanned bilingual contracts all work.
One number, 0-100, summarising your contract's overall risk position.
Upload a ZIP of contracts and ask one question across all of them.
What-if engines for litigation and contracts — outcome odds, damages, settlement, and AI negotiation roleplay.