Designing for AI explainability: Briefly's assistant UX redesign balances speed and transparency for lawyers

AI · 6 min read

Briefly provides AI-assisted brief drafting for law firms and faced an adoption bottleneck: lawyers were skeptical of black-box outputs and spent significant time cross-checking citations and reasoning. The product team needed to balance the friction of verification with the time-savings promise of AI. They launched a redesign to embed lightweight explainability into the assistant UX.

The redesign introduced a ‘why this suggestion’ affordance that surfaced concise provenance (source snippets, confidence bands, and a linked audit trail) beneath each generated paragraph. It also added an inline citation validator that flagged low-confidence references and provided one-click access to original documents. Crucially, the team kept the assistant fast: explanations were asynchronous and could be expanded on demand to avoid interrupting drafting flow.

In pilot tests with three mid-size firms over eight weeks, perceived trust in AI outputs increased 23% and the average time lawyers spent verifying AI suggestions dropped by 39%. The audit trail also reduced risk concerns for compliance teams. Briefly's case shows that explainability in product UX is not just an ethics checkbox but a conversion lever — when designed to be minimally intrusive and contextually available, it accelerates adoption among skeptical professional users.