Overview

Automated legal document review

As part of the 2-day yuuvis hackathon, I worked with a team to create a product that incorporated content management with the yuuvis API.

Technology in the legal space is extremely outdated, so we set out to create a way for lawyers to navigate massive amounts of legal documentation. By implementing a chatbot that streamlines and automates legal document review and segmentation, we will save lawyers time and help them win cases.

Problem

Immigration lawyers have extremely limited resources

Immigration lawyers are some of the most over-worked and underpaid in the legal industry. It’s no secret that lawyers and legal assistants spend countless hours pouring through past case documentation to establish a legal precedent. The current immigration crisis means the demand for legal representation for immigrants is higher than ever. Even saving the immigration lawyer 1 hour could be the difference of whether or not someone gets deported.

Users & Audience

Lawyers are the primary users of this tool

As we did not have time to do a ton of research, I reviewed profiles online for immigration lawyers in Austin to create a proto-persona. This helped us craft the narrative as a team of strangers to unify what we working on and who it was for. Additionally, it helped us as we sought to explain the use case of our product to a room of 80+ people.

Roles & Responsibilities

• Timeline: 2 day hackathon

• Team:

  • Liberty Gallagher (me) - UX Designer, Project Manager

  • Xlegic Sin’Austin - Data Scientist, Team Lead

  • Jacob Schaefer - Full Stack Developer

  • Jeff Pape - Back-end Developer

  • Alan Ngo - Back-end Developer

• Tools:

  • Design - Figma

  • Development - JavaScript, Dialogflow, Node.js, Java, Python, TensorFlow, yuuvis API

Design Process

Planning

Since the team was all working simultaneously, we mapped out milestones for each function to understand what each other needed and was working on.

Sitemap

Touch Points

Since legal tech is not the most advanced, we wanted to create multiple ways for lawyers to access this chatbot regardless of the technology they have access to. Not all of them would be willing to use a brand new interface, so we integrated into other systems they are already familiar with. In total, we created five different points of access for a lawyer to interact with the bot. Three of the touch points originate on the web interface through the search functionality.

Early Sketches

To get a feel of what the interface needed to look like, I collaborated with the Data Scientist (and former legal assistant) to understand what type of information we could pull from the database.

Prototype

Next Steps

We accomplished so much as a team during the hackathon. However, there’s a lot more we could do with this project. Looking into the future we would want to:

• Research and test - We could craft a more accurate persona and a better understanding of our users if we interviewed immigration lawyers. Usability testing with real lawyers would help us understand current challenges with the UI and allow us to prioritize the most important features 

• Build out front-end - Our developers didn’t have time to code the UI design during the hackathon. After some testing and iterations, we’d love to build out the actual interface so users are not just limited to the web plugin and slack. 

• Add education (on slack/web plugin) - These methods put more responsibility on the user since it’s free-text based. To make sure these methods are as successful as possible, we need to implement some basic education about the types of things they can ask the bot so their interactions are as productive as possible.