Lawyers using Anycase.ai's AI-powered research struggled to trust AI-generated answers. While the AI provided accurate information, lawyers needed to verify every citation—their reputation depends on accuracy. The existing interface treated citations as afterthoughts (simple blue links), ignoring how lawyers actually work with legal references.
Anchor the interface to what is familiar to user
The design was anchored on observations on how lawyers read and analyze files: through their trusty highlighters. Not only is this familiar, but it is a great break from such long and boring text. They read the paper once, but then most of the time, their intent is to scan and look for something specific. This is the first step into building trust.
Citation as Highlights
Hover Preview System
Sidebar Reader Integration
The redesigned interface transforms how lawyers interact with AI-generated legal research. Citations now appear as highlighted tabs (mimicking lawyers' annotation habits) rather than simple links, with color-coding to distinguish laws (yellow), jurisprudence (blue), and regulations (green). On hover, a popover displays the actual canonical document preview—not just metadata—allowing instant verification without losing context. The sidebar reader provides a focused view where lawyers can dive deeper, with a "copy to cite" feature for easy reference building. The entire interface shifted to serif typography and traditional legal document styling, making the AI feel familiar rather than foreign. This multi-layered approach—from quick hover previews to deep-dive reading—respects how lawyers naturally verify information while building trust through transparency.
Sometimes innovation means making new technology feel comfortably familiar. Trust in AI must be supplemented with good user experience through familiar patterns and transparent verification.
Trust Through Citations: Redesigning AI Legal Research for Litigation Lawyers