How Eleken Helped CaseScribe Launch a User-Ready Legal AI Platform

How Eleken Approached the Challenge

During a 3-day trial and continued collaboration, Eleken’s designers focused on transforming CaseScribe’s MVP from a fragmented interface into a platform that supported real legal workflows.

The biggest problem we uncovered during the trial was that documents and their extracted facts were split across different tabs. For lawyers, this meant constantly switching screens, losing context, and struggling to connect documents with their key case details.

Eleken redesigned this workflow into a unified view:

• Facts remained visible on the left side of the screen.
  
• Documents opened on the right, so users could review them without losing context.
  
• Clicking a document automatically filtered the facts table, showing only the relevant entries.

This change turned a fragmented, frustrating process into a smooth and contextual review experience: a single screen where lawyers could read, analyze, and act. On top of that, we introduced sorting options for the facts table and the ability to add comments, giving users more control and flexibility in their daily casework.

By the end of the trial, CaseScribe didn’t just have cosmetic improvements. They had a scalable design direction that allowed them to onboard their first real users and start collecting feedback, while also setting the stage for AI-powered functionality.

UX Audit as a Starting Point

Eleken kicked off the long-term collaboration with a UX audit of CaseScribe’s MVP. The goal was to identify where users struggled most and to prioritize improvements that would make the product usable for real legal teams.

The audit revealed that many challenges went deeper than visuals. They were rooted in workflows, structure, and consistency:

• Technical gaps: Users had no clear feedback when the system was processing documents, leaving them unsure if tasks were running. Navigation also wasn’t scalable enough to handle future AI features and new product areas like “Clients” or “Research.”
• Inconsistent patterns: Buttons and interactions behaved differently across screens, and small glitches (like scrolling buttons) undermined trust.
• Layout and hierarchy issues: Key elements were separated across tabs, with documents in one view and facts in another, making it hard to connect them. Poor hierarchy further reduced clarity, leaving users unsure what to focus on.
• UI problems: The interface relied on tiny PDF previews, lacked modal confirmations for critical actions, and overall felt outdated and inconsistent.‍
• Forms and input friction: Data entry didn’t guide users clearly, creating frustration during essential workflows.

Instead of just restyling screens, the audit showed the need for a broader UX rethink: unify documents and facts, fix inconsistencies, and create a foundation for scalability. This became the blueprint for Eleken’s design work.

Fixing Critical UX Gaps

When Eleken joined, CaseScribe’s MVP had the right AI engine but the wrong experience around it. Core workflows like reading documents, reviewing facts, and filtering information were clunky and fragmented. Our job was to remove friction, connect documents with their data, and give legal teams an interface they could rely on.

Fixing the PDF Viewer

One of the biggest blockers in CaseScribe’s MVP was the Document View. The hierarchy was confusing: 

• a back arrow that didn’t behave consistently, 
• a document list hidden in odd places,
• a preview window so small it was almost unusable. 

Lawyers and paralegals couldn’t comfortably read files or keep track of where they were.

Eleken restructured the viewer into a clear, three-part layout: a list of documents on one side, the selected document displayed at a readable size in the center, and its related facts linked directly within the same view. This solved hierarchy issues and gave users context and control in one place. No more switching back and forth.

Making Facts Usable

Another major pain point in CaseScribe’s MVP was how Facts were displayed. They were presented in long, compressed tables with no spacing, no grouping, and no clear way to trace them back to the original document. Users complained the screen “blurred in the eyes,” and they often lost context when trying to review information.

We redesigned the tab into a document-first hierarchy: each file at the top, followed by its extracted facts. Users can group facts by tags or dates, depending on their needs. To keep context, the related PDF is always visible on the right side. This way, facts aren’t just raw data; they’re structured, easy to scan, and directly tied to their source.

This redesigned flow not only solved immediate usability issues but also inspired the product’s future redesign and helped CaseScribe secure a major investor.

Expanding AI Capabilities

CaseScribe’s MVP started with basic AI functionality, but it wasn’t enough to support real legal workflows. Together with Eleken, the platform began evolving into a more powerful, user-ready tool with new AI-powered features:

1. AI extracts facts

One of the most important upgrades was making AI-extracted facts usable in real workflows. We unified the UI so the Documents tab mirrored the Facts tab, creating a predictable, consistent experience. Within this flow, AI automatically extracted facts from documents and displayed them directly alongside their sources. 

This way, lawyers could read a document and immediately see its structured facts without switching tabs or losing context. The result was a smoother, faster review process that made the AI’s work transparent and reliable.

2. Automatic tag summaries 

We applied AI-generated tags directly to facts, enabling lawyers to group data by conditions, diagnoses, or timelines. The AI then summarized multiple tagged facts into one short, meaningful overview. For example, instead of filtering through five different entries about “bilateral hearing loss,” a lawyer could click the tag and instantly read a concise summary of the client’s medical history.

3. AI-powered chatbot (in progress) 

To make interaction even more natural, Eleken worked on designing an AI assistant that allows users to query documents and facts directly. The chatbot is envisioned in two modes:‍

• ‍General chatbot – Users can ask case-level questions, and the assistant analyzes all available documents. It not only answers queries but also references the documents it drew from, suggests next steps, and can even format output into reports ready to export.

• Filter and tag chatbot

Integrated into the Facts and Case Summary tabs, this assistant adapts to user context. In the Facts tab, it works with applied filters or tags, narrowing its analysis to provide focused insights with tailored prompts.

In the Case Summary tab, users can launch the chatbot directly from a specific tag (e.g., “left knee”), and it surfaces insights only from facts linked to that tag, enabling highly precise case analysis.

From MVP to Market-Ready

With Eleken’s help, CaseScribe transformed its MVP from a clunky proof of concept into a user-ready platform that lawyers and paralegals could actually rely on. By fixing core UX issues, like the broken PDF viewer and scattered fact tables, we created smoother workflows that matched real legal practices.

At the same time, we laid the foundation for AI-driven growth, moving beyond simple summarization to features like automatic tagging, consistent navigation, and early chatbot prototypes. The redesign not only improved daily usability but also inspired CaseScribe’s future product vision and helped secure a new investor.