ai powered research
With the release of GPT-3.5 in November 2022, we aimed to build an AI-first web app for researching stocks, replicating the interactive depth and subtlety of conversations with a human analyst. Users could interact with the AI, asking questions about companies as if speaking to the CEO, and receive the AI's stance on stock prices, derived from a wealth of information.
The app featured a real-time knowledge graph displaying qualitative and quantitative data about companies from sources like investor sites, earnings calls, social media, and news. Inspired by the Bloomberg terminal, the design was familiar to most users, maintaining a similar style.
DYNAMIC CHAT INTERFACE
early prototype to show DASHBOARD & INTERACTION
Development of the MVP
The idea behind the design was to create a white-label product that clients could customize with their brand colors. We hired a freelance engineer and worked with him to build an MVP within 5-6 weeks. The decision on the MVP was based on technical limitations and what could be built quickly and cheaply so we could test it effectively.
AI PERSONA - ralina
Ralina responds to your questions and provides sourcing material, indicating the specific PDF and page number for easy reference. Clicking on the source gives quick access to the relevant page and document. We added fun elements to make Ralina feel "alive," like blinking and showing anger if you tap her face too much. We had plans for a range of emotions based on tone and questions, but implementing these proved too technically challenging.