Emmi: AI Legal Assistant Mobile App by Ella Sh.Emmi: AI Legal Assistant Mobile App by Ella Sh.

Emmi: AI Legal Assistant Mobile App

Ella Sh.

Ella Sh.

Emmi is an AI assistant built specifically for legal professionals. A single chat window handles client onboarding, risk assessments, document generation, and practice area guidance.
The problem. Legal work runs on precision and time. Neither one survives contact with a clunky internal tool. The harder design problem was trust: an AI handling sensitive, high stakes information has to read as a reliable colleague rather than a search engine, and it has to do that in the first ten seconds.
Structure instead of a blank prompt. The home screen organises the assistant's capabilities into clearly labelled practice areas: Onboarding, Intellectual Property, Bankruptcy, Employment, Civil Litigation, Real Estate. Users start from a category rather than an empty input. That single decision removes the blank page problem and hands the model its context before the first message is typed.
The conversation carries the work. Emmi guides users through structured workflows. It asks one question at a time, confirms each action with a visible checkmark, and surfaces downloadable documents and Clio integrations directly inside the chat. Nothing requires leaving the conversation to complete a task.
Visual register. Deliberately restrained. A warm off white base, muted pastels for the category icons, and a hand lettered wordmark give the product personality without undermining the professional tone. Clean typography and generous spacing keep the interface readable across long working sessions.
Result. An AI tool that fits naturally into a legal workflow. Structured enough to be trusted, approachable enough to open every day.
Timeline: 5 weeks. Year: 2024. I led the product design end to end, from UX architecture through final UI.
Like this project

Posted Aug 28, 2026

UX/UI design for an AI assistant built for legal professionals. Client onboarding, risk assessments, and document generation inside one conversational interface.