Artificial Intelligence Data Extraction Platform
Re-engineered an app that let clients view, search, and generate reports based off their contract document library.
For this project, I needed to fix issues with creating reports, searching data, as well as locating and viewing contracts or specific contract data. Our customers used the platform as a way to make business decisions revolving around expiring lease contracts, unique clauses relating to terminations or renewals, revenue numbers and loses, among other business critical data.
Platform Triage
To begin with, I studied research material in our tech library. I wanted to see what was lacking as far as relevant data to begin building a plan of action. Key to this was understanding user workflows. Users had difficulty locating data due to the inefficiency of the search tool. I found this was in relation to the AI data model structure and users misunderstanding of it.
As well, the app was overly complex due to the complex nature of the data models. Client contracts could be voluminous, containing around 800 individual data points or more.
There were a number of different user personas for the app, from basic legal data clerks who would search contracts for names, clauses, or specific details, as well as senior management requiring high level overviews, who would leverage the entire data model to generate reports leading to large business decisions.
OOUX and Workflows
Together with my team, we began to make an object map of the app to understand the components involved. This would help generate an inventory to build the site map and Information Architecture. As well this would help define the scope of the work load.
*Proprietary data has been censored
Feature Prioritization
We then created a prioritization framework to assign values and weight to features, which customer’s prioritize and value. This would also help prioritize development work and sprint planning.
*Proprietary data has been censored
Design Sprints and Wireframing
After we had some direction, I organized some design sprints with stake holders to come up with new solutions. We worked with the Google Design Sprint concept to try and come up with solutions and new ideas.
Prototyping and User Tests
Once I had some ideas down digitally, I scheduled interviews with stakeholders and clients. I began to get feedback based on our ideas, and then make changes to the prototypes based on all the feedback I received in testing. Then I went back to those same clients, as well as new clients to see if the changes were successful, and gather more feedback.
Final Designs for Development
After user testing had yielded positive results, I created new visuals and UI components to finalize designs. I coordinated with developers to being creating tickets to being sprint planning.
Data Analytics
The final task was to create an analytics app to integrate with the searching and filtering of client data. This would be used to generate reports and analytics for sharing across teams and staff, as well as externally for presentations or meetings. This was a front end UI that would connect APIs between Leverton and third party systems clients would be using.