Screen from a fictional news mobile app on a blue gradient background
Screen from a fictional news mobile app on a blue gradient background

Generative AI

RESEARCH

RESEARCH

BRANDING

BRANDING

SidebarAI – Designing Filevine’s First AI Assistant
Lead UX Designer | Spring 2023

Overview

In early 2023, Filevine set out to launch its first generative AI feature: a conversational assistant built into the Sidebar. I led this project as the sole designer on a cross-functional team, working closely with a PM, three engineers, and two QA testers. The goal wasn’t just to ship something AI-flavored — it was to solve a real, painful problem: users couldn’t find what they needed inside Filevine.

Why this problem?

Filevine is a robust case management platform for legal professionals — think of it as the operating system for law firms. But with power came complexity. Users (especially legal assistants and lawyers under pressure) constantly struggled with discoverability. They’d often turn to Slack, Google Drive, or colleagues just to figure out how to use the tool they were paying for.

This wasn’t just a UX problem — it was a business risk. If users can’t find key features, they churn or underutilize the platform. This project aimed to reverse that trend and position Filevine as a leader in legal tech innovation.

Business Goals

  • Launch Filevine’s first generative AI feature ahead of Lex Summit, our annual product showcase

  • Improve feature and information discoverability across the product

  • Lay the groundwork for future AI development within the platform

  • (Estimated long-term impact: increased feature adoption + reduced support tickets = ~$X savings annually)

The MVP Constraints

To start, I partnered with engineering to understand technical feasibility. For the MVP, the assistant would only be trained on internal Zendesk and Helpdesk content — so it could handle “how do I…” style support queries, but not deeper legal case content.

Early Design + User Research

I started with a basic chat UI mockup and ran early feedback sessions to understand how users might engage with an in-app AI tool. Two key user types emerged:

  • Tech-savvy users: Already familiar with AI and excited to test
  • Non-tech-savvy users: Unsure what to ask or how this would help them

The blank text field created a cognitive block. Users didn’t know where to begin.

Key Insight → Solution: Suggested Prompts

To unblock users, I introduced Suggested Prompts — a list of click-to-start questions that appeared via a lightning icon. These served as conversation starters, helping even hesitant users explore the AI's capabilities. For the MVP, prompts were pulled directly from common ticket themes.

AI Branding + Design Foundations

Beyond the MVP, I established AI design standards for Filevine:

  • Iconography (selected a star symbol — now widely associated with AI)

  • A cohesive, scalable SidebarAI component system

  • Tone and behavior guidelines for future conversational UIs

This work set the tone for every AI feature Filevine would build next.

Outcomes & Results

  • Launched SidebarAI in beta ahead of Lex Summit

  • 70% of users interacted with Suggested Prompts in the first 2 weeks

  • Stakeholder excitement led to accelerated roadmap for AI features

  • Laid scalable design foundations for AI across Filevine

Abstract image used as a placeholder for this design project
Abstract image used as a placeholder for this design project
Abstract image used as a placeholder for this design project
Abstract image used as a placeholder for this design project
Abstract image used as a placeholder for this design project
Abstract image used as a placeholder for this design project