Senior Product / UX Designer
SF Bay Area
hi@mark-b.com
// Case study

LexisNexis Lexis+

Legal research & analytics

Applied AI concepts

2016–22

In 2016, LexisNexis asked what legal research would look like when the system could think. Almost a decade later, the answers are the product.

Role
Design lead for the AI research vision, the conversational assistant workstream, and the litigation-analytics integration, across an eight-year client relationship.
Team
Awasu Design studio team, embedded with LexisNexis product, UX research, and iLabs data science; later with Lex Machina product and engineering.
Client
LexisNexis (Lexis Advance, then Lexis+) and Lex Machina.
Year
2016 to 2022. Won on the strength of earlier Lex Machina work, and ended with designs shipping inside Lexis+.
Conversational Research Assistant Citation Visualization Litigation Analytics An Eight-Year Client Relationship

The problem

The deepest library in the world, behind a search box that assumed you were an expert

Content this deep is only as useful as a lawyer's confidence in what they have found. Users told LexisNexis the same four things. The design problem was to make the intelligence helpful without making the interface harder.

I don't know how to query.

I don't know where to start.

I don't know how long this will take.

I don't know when I'm done.

Approach

Put the intelligence behind the interface, not on it

Simple in the front, smart in the back became the engagement's operating principle and, over eight years, its through-line: a two-direction vision system, a research-validated conversational assistant, the citation visualization's entry into the core product, the analytics line's naming, and the litigation-analytics module inside Lexis+.

Arc timeline, 2016 to 2022, five engagement markers · build fresh · NDA low

Act 1 · 2016

A ten-week vision system, and a work group formed on the call

Four user questions as the frame, personas and a real scenario, then roughly 96 screens across two deliberately staged directions: a near-term distributed intelligence, and a future-state assistant. An InVision prototype and HTML specs backed a 90-minute executive presentation. The room split, and that split set up every later act: a regression worry on the first direction, and real enthusiasm for the way the second answered navigation.

Principle slide · simple in the front, smart in the back · Final Review p14 · NDA low
D1 vs D2 split-screen · Final Review p23 · light scrub · NDA med

The object model

Every object in the product, and every path between them

Before any screens, I mapped the platform's object model: every object Lexis Advance held, and every route a researcher could take between them. It anchored the strategy sessions and opened the executive presentation. It also outlasted the engagement. A year later two concurrent workstreams ran on the same map, the same persona and the same journey, which is how the anti-silo mandate from that spring's workshop got answered in something checkable.

The 2016 Lexis Advance object model: thirty-five product objects arranged on
                      and around the home surface, the paths a researcher could take between them,
                      and the actions available on each object.

Act 2 · 2017

Query formulation was the threshold problem, so that is where the assistant started

Seven problem-discovery interviews surfaced two things: lawyers could not formulate the query, and they could not tell when they were finished. A journey map, a 21-concept Guided Research system, and chatbot concepts went to leadership, then the work pivoted to query assist. Prototypes were tested with a real firm, co-designed with the client's product lead in shared files, against two MVP principles: guide, don't replicate, and nail it before you scale it.

The 21-concept one-pager · full-page artefact scan, scrubbed · NDA low · BEST single overview of this act
Conversational flow strip, 3-4 frames · the 3,140 to 240 narrowing · Chatbot Concepts pp34-44 · NDA med, scrub
Journey map · wide artefact scan or clean redraw · NDA low

Sidebar · Shepard's

How an acquired visualization entered the Shepard's experience

Our studio designed the way the acquired citation visualization was folded into Shepard's. It launched publicly as Ravel View in 2018.

Sparkle sparkline concept · NDA med, scrub
Concept
Ravel View · public 2018 press screenshot · credit
Shipped, 2018

Act 3 · 2018

Naming the analytics line, and a term visualization that fed the assistant

Post-acquisition brand architecture with a naming partner, against two criteria: coexistence with Lex Machina (broad alongside deep), and confidence positioning. Facilitated the brand architecture and naming for the analytics line that launched as Lexis Analytics.

QFL visualization sketches, 2-3 · ABC deck pp4-10 · NEEDS the two-page scrub fix first · NDA low after

Act 4 · 2020–22

Designing the litigation analytics module inside the new platform

Embedded with Lex Machina and LexisNexis to design the Litigation Analytics module for Lexis+, on a use-case framework of entities against jobs, with dual data sources and gateway positioning. Announced January 2021. Then the state-court era: document retrieval, Motion Mapping, and a roughly 500-hour Cross-Court State Search system that was part logic puzzle and part UI. Handed off October 2022, demoed shipped in 2023.

Use-case matrix, entities x jobs · clean diagram REDRAW · NDA low redrawn
Lexis+ Litigation Analytics · public product screenshot · credit
Cross-court selection flow · 2-3 scrubbed frames or redrawn diagram · NDA med scrubbed

The lineage

What shipped where

Concept on the left, shipped feature on the right, with the connector weighted by how strong the claim is. Verbs are calibrated and should not be upgraded.

Conversational assistant concepts 2017

groundwork

Lexis+ AI conversational search 2023

An intelligent identity 2016

pattern first

Protege personalized assistant 2025

Citation visualization integration 2017

designed

Ravel View + Shepard's case cards 2018

Analytics naming 2018

designed

Context, in Lexis Analytics 2018

Analytics module design 2020

designed

Lexis+ Litigation Analytics 2021

Motion Mapping + Cross-Court Search 2021

designed

State Motion Metrics + state-court search 2022

Case-facts search concept 2017

pattern first

Fact & Issue Finder 2022

Document intelligence concepts 2017

pattern first

Brief Analysis 2020

Word / Teams concepts 2019

pattern first

Lexis Create, Protege for Teams 2021

Solid: designed Dashed: groundwork Dotted: pattern first
8Years
10+Workstreams
LiveInside Lexis+
2016The bet

The concepts became the product.

This is awesome. We are very much on the same page!
Mac Taylor, LexisNexis
I'm so glad we brought Awasu into this. This is the kind of challenge we need you for.
Wade Malone, LexisNexis
PLACEHOLDER: Serena's process quote, pull the exact line from the engagement-story note.
Serena, LexisNexis