DocuTrust

DocuTrust is a concept for AI-assisted contract review — software that reads through a contract, flags risky language, explains the risk in plain terms, and helps legal and procurement teams get from a first read to a decision faster. The homepage leads with a real-feeling example: a Master Services Agreement clause flagged "High Risk" for an uncapped liability limitation.

Legal-tech products live or die on whether the reasoning behind a flag feels trustworthy. The challenge was designing an interface that explains *why* a clause is risky — not just that it is — without ever implying the tool replaces a lawyer's judgment.

Give every document a legible state — verified, pending, rejected, expired — so trust in the system doesn't depend on reading fine print.

I built the signature interaction around the redline itself: a draggable before/after comparison switching between the original clause, the AI-suggested redline, and a clean final version, paired with a document-level risk breakdown (liability, termination, indemnification, data, payment) so a visitor can see both the micro (one clause) and macro (whole contract) view of risk.

A dark, serif-forward interface that reads more like a legal workspace than a typical SaaS dashboard, with a single amber accent reserved specifically for risk flags so it never gets diluted by decorative use elsewhere.

Built with semantic HTML5, modular CSS and vanilla JavaScript. The redline comparison, risk scoring and review workspace all run on a fixed illustrative contract example — the site is explicit that DocuTrust assists review but does not replace qualified legal counsel, and holds no third-party security certifications.

A document-technology concept that demonstrates how AI-assisted contract review could explain its reasoning clause by clause — not just flag risk — while being explicit that it assists rather than replaces professional legal review.

State design is the real interface design problem in document products — most of the work is deciding what a document looks like when something has gone wrong.