Simplifying the work of tax professionals with a mobile app
Industry
Legal services
Output
Application
Timeline
Fall to Winter 2016
Office
Rangle Toronto
Highlights
- Agile delivery of an AI-powered tax research tool from concept to product
- App architecture designed to scale additional AI-powered services easily
- Enabled client to take future development in-house with collaborative work and seamless knowledge transfer
The challenge
Blue J Legal's vision was to simplify the process of decision-making in tax research and planning. Their goal was to build an AI-powered research tool that professionals could use to ask a complex tax law question and receive a definitive answer, backed up by relevant research articles.
As a pre-revenue startup with a small team, Blue J’s challenge was to transform their idea from a heroku-based mockup to a fully-deployed product as quickly as possible. To make the app a reality, they partnered with Rangle, combining their legal expertise with our skill in React development, machine learning, and team enablement.
As a pre-revenue startup with a small team, Blue J’s challenge was to transform their idea from a heroku-based mockup to a fully-deployed product as quickly as possible. To make the app a reality, they partnered with Rangle, combining their legal expertise with our skill in React development, machine learning, and team enablement.
The process
Rangle’s team worked with Blue J to deliver a viable product on a short timeline. The app was a React frontend and node.js-based backend architecture that integrated with IBM Watson and Microsoft Azure to provide tax classifications and source the top five case studies supporting the result. The team delivered on an incredible user experience that remains the foundation of Blue J’s AI-powered services even to this day.
The Rangle team also ensured longevity, creating a platform that could easily scale the launch of additional machine learning services over time. This included a Classifier Generator that reduced the build time needed to add additional classifiers from one month to a single week, allowing Blue J to rapidly increase the scope of tax-law questions their platform could answer.
Powered by machine learning
App users fill out a questionnaire during each new query session, which results in a classification on a legal tax question (e.g. whether a particular expense should be categorized as “Tangible” or “Intangible”), a confidence rating, and supporting cases. Users email the results to themselves, and provide feedback on the session and the app. Results are generated by querying Machine Learning services (eg, IBM’s Watson and Microsoft’s Azure). These services are trained by Blue J’s tax law experts to provide accurate classification, as well as context for classifications that aren’t as straightforward.
The outcome
Blue J Legal successfully launched its MVP and acquired its first paying customers during their partnership with Rangle. The team enablement provided by Rangle helped them expand and level-up their staff to take app production in-house. Blue J Legal now offers a suite of over 40 classifiers, and has since raised more than $9M in venture financing.